sg245280
Beyond DHCP Work Your TCP/IP Internetwork with Dynamic IP
Oscar Cepeda, Bob Chambers, Julian Mosca, Matt Robbins
International Technical Support Organization
www.redbooks.ibm.com
SG24-5280-01
SG24-5280-01
International Technical Support Organization
Beyond DHCP Work Your TCP/IP Internetwork with Dynamic IP
January 2000
Take Note!
Before using this information and the product it supports, be sure to read the general information in
Appendix E, “Special notices” on page 485.
Second Edition (January 2000)
This edition applies to AIX Version 4.3, Windows NT Server, OS/2 Warp Server for e-business, OS/2
Warp Server Version 4, OS/2 Warp 4, Workspace On-Demand, and TCP/IP Version 4.1 for OS/2 Warp.
This edition also applies to Lotus Notes, Lotus Domino Go Webserver, and Windows 3.1, Windows 95
and Windows NT client enhancements from IBM.
Comments may be addressed to:
IBM Corporation, International Technical Support Organization
Dept. DHHB Building 003 Internal Zip 2834
11400 Burnet Road
Austin, Texas 78758-3493
When you send information to IBM, you grant IBM a non-exclusive right to use or distribute the
information in any way it believes appropriate without incurring any obligation to you.
© Copyright International Business Machines Corporation 1998, 2000. All rights reserved.
Note to U.S Government Users – Documentation related to restricted rights – Use, duplication or disclosure is
subject to restrictions set forth in GSA ADP Schedule Contract with IBM Corp.
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi
The team that wrote this redbook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Comments welcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi
Chapter 1. TCP/IP overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 Network protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 IP addresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 IP subnets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4 IP routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5 Assigning IP addresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5.1 BOOTP and the Dynamic Host Configuration Protocol (DHCP)
1.6 Domain Name Services (DNS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.6.1 Domain names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.6.2 Domain versus zone of authority . . . . . . . . . . . . . . . . . . . . . . .
1.6.3 Differentiating name servers . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.6.4 Dynamic Domain Name Services (DDNS) . . . . . . . . . . . . . . . .
1.6.5 Resource record types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.7 Network security and the Internet . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.8 IP Version 6 (IPv6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.9 Applications that use TCP/IP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.10 Other TCP/IP terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. .1
. .1
. .2
. .6
. .9
. 10
. 10
. 12
. 13
. 16
. 18
. 22
. 23
. 25
. 26
. 26
. 26
Chapter 2. Up and Running with DHCP. . . . . . . . . . . .
2.1 Why DHCP? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 OS/2 Warp Server for e-business as a DHCP server
2.2.1 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.2 DHCP server preparation . . . . . . . . . . . . . . . . .
2.2.3 Basic configuration . . . . . . . . . . . . . . . . . . . . . .
2.2.4 Testing and verification. . . . . . . . . . . . . . . . . . .
2.3 Windows NT as a DHCP server . . . . . . . . . . . . . . . .
2.3.1 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.2 DHCP server preparation . . . . . . . . . . . . . . . . .
2.3.3 Basic configuration . . . . . . . . . . . . . . . . . . . . . .
2.3.4 Testing and verification. . . . . . . . . . . . . . . . . . .
2.4 AIX as a DHCP server . . . . . . . . . . . . . . . . . . . . . . .
2.4.1 Manually configuring the DHCP server . . . . . . .
2.4.2 Configuring DHCP server through dhcpsconf . .
2.4.3 Disabling BOOTP support in inetd . . . . . . . . . .
2.5 DHCP clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.1 OS/2 Warp 4. . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.2 Windows 95 and Windows 98 . . . . . . . . . . . . . .
. 29
. 30
. 30
. 30
. 31
. 35
. 40
. 43
. 43
. 44
. 45
. 49
. 50
. 50
. 51
. 62
. 64
. 64
. 72
© Copyright IBM Corp. 1998, 2000
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
iii
2.5.3 Windows NT Workstation 4.0 . . . . . . . . . . . . . . . . .
2.5.4 Windows 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.5 Apple Macintosh . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.6 IBM Network Station . . . . . . . . . . . . . . . . . . . . . . . .
2.5.7 AIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.8 Linux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.9 DOS, Windows 3.1, and Windows for Workgroups .
2.5.10 Sun Microsystems Solaris. . . . . . . . . . . . . . . . . . .
2.5.11 IBM WorkSpace On-Demand . . . . . . . . . . . . . . . .
2.5.12 Hewlett-Packard LaserJet 4000 Printer (JetDirect)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. . 73
. . 75
. . 77
. . 79
. . 86
. . 90
. . 92
. . 99
. 102
. 105
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 113
. 113
. 113
. 114
. 114
. 115
. 115
. 116
. 116
. 117
. 118
. 118
. 118
. 120
. 122
. 126
. 127
. 128
. 128
. 129
Chapter 4. Serving names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1 Name serving on the Windows platforms . . . . . . . . . . . . . . . . . .
4.1.1 Windows NT as a Static DNS server. . . . . . . . . . . . . . . . . .
4.1.2 Windows NT as a DDNS using WINS . . . . . . . . . . . . . . . . .
4.1.3 Windows 95/98 as a static DNS client. . . . . . . . . . . . . . . . .
4.1.4 Windows NT Workstation 4.0 as a static DNS client . . . . . .
4.1.5 Windows with IBM Dynamic IP Client . . . . . . . . . . . . . . . . .
4.2 OS/2 Warp Server as Dynamic DNS server . . . . . . . . . . . . . . . .
4.2.1 The Dynamic Domain Name System (DDNS) . . . . . . . . . . .
4.2.2 The IBM Dynamic DNS (relative to other implementations).
4.2.3 Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 131
. 131
. 131
. 136
. 142
. 142
. 143
. 150
. 150
. 156
. 159
Chapter 3. DHCP server and client interaction .
3.1 DHCP basics . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1 Client states . . . . . . . . . . . . . . . . . . . . . .
3.2 The DHCP lease . . . . . . . . . . . . . . . . . . . . . .
3.2.1 Getting the lease . . . . . . . . . . . . . . . . . .
3.2.2 Renewing the lease . . . . . . . . . . . . . . . .
3.2.3 Rebinding the lease . . . . . . . . . . . . . . . .
3.2.4 Eviction . . . . . . . . . . . . . . . . . . . . . . . . .
3.3 Interaction between DHCP server and client .
3.3.1 BOOTP. . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.2 Relay agents . . . . . . . . . . . . . . . . . . . . .
3.4 DHCP messages . . . . . . . . . . . . . . . . . . . . . .
3.4.1 DHCPDISCOVER . . . . . . . . . . . . . . . . . .
3.4.2 DHCPOFFER . . . . . . . . . . . . . . . . . . . . .
3.4.3 DHCPREQUEST . . . . . . . . . . . . . . . . . .
3.4.4 DHCPNAK . . . . . . . . . . . . . . . . . . . . . . .
3.4.5 DHCPACK . . . . . . . . . . . . . . . . . . . . . . .
3.4.6 DHCPDECLINE . . . . . . . . . . . . . . . . . . .
3.4.7 DHCPRELEASE . . . . . . . . . . . . . . . . . . .
3.4.8 DHCPINFORM . . . . . . . . . . . . . . . . . . . .
iv
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
4.2.4 Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.5 Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.6 Setting up your DDNS server . . . . . . . . . . . . . .
4.2.7 ProxyArec. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.8 Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.9 DDNS files . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.10 Providing mail services . . . . . . . . . . . . . . . . . .
4.3 Dynamic DNS client support in OS/2 Warp Server . .
4.3.1 OS/2 Warp 4 with TCP/IP Version 4.1 for OS/2
4.3.2 OS/2 Warp 4 with TCP/IP 4.0 . . . . . . . . . . . . . .
4.3.3 OS/2 Warp 4 using presecured mode . . . . . . . .
4.3.4 WorkSpace On-Demand using proxy . . . . . . . .
4.4 Name serving on AIX . . . . . . . . . . . . . . . . . . . . . . . .
4.4.1 AIX as a DNS client . . . . . . . . . . . . . . . . . . . . .
4.4.2 AIX as a static DNS server . . . . . . . . . . . . . . . .
4.4.3 AIX as a DDNS server . . . . . . . . . . . . . . . . . . .
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 160
. 160
. 161
. 176
. 177
. 177
. 185
. 198
. 198
. 201
. 203
. 206
. 208
. 209
. 211
. 218
Chapter 5. Integrating file and print services . . . . . . . . . . . . . . . .
5.1 NetBIOS, NetBEUI and TCPBEUI: What are they? . . . . . . . . . . .
5.2 Resolving NetBIOS names to IP addresses . . . . . . . . . . . . . . . .
5.2.1 Enhancing B-Node clients . . . . . . . . . . . . . . . . . . . . . . . . .
5.3 New way of using DNS with non-RFC-encoded name . . . . . . . .
5.4 NetBIOS Name Servers (NBNS). . . . . . . . . . . . . . . . . . . . . . . . .
5.4.1 Microsoft WINS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.2 Network TeleSystems Shadow IPserver . . . . . . . . . . . . . . .
5.5 P-Node, H-Node and M-Node clients . . . . . . . . . . . . . . . . . . . . .
5.5.1 Manually configuring clients to use an NBNS . . . . . . . . . . .
5.5.2 Dynamically configuring clients to use a NBNS. . . . . . . . . .
5.6 IBM Neighborhood Browser Enabler for Warp Server . . . . . . . . .
5.6.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.6.2 Installing neighborhood browser enabler . . . . . . . . . . . . . .
5.6.3 Starting and stopping the Neighborhood Browser Enabler .
5.7 AIX Fast Connect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.7.1 Fast Connect on static server with dynamic client. . . . . . . .
5.7.2 Fast Connect on a dynamic server with dynamic client . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 227
. 227
. 232
. 233
. 246
. 253
. 253
. 255
. 257
. 257
. 262
. 271
. 271
. 272
. 274
. 274
. 274
. 275
Chapter 6. Expanding your network . . . . . . . . . . . .
6.1 Multiple subnets . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.1 Static and dynamic routing between subnets
6.1.2 OS/2 Warp as a router . . . . . . . . . . . . . . . . .
6.1.3 Windows NT as a router . . . . . . . . . . . . . . . .
6.2 Crossing routers . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.1 Windows NT as a DHCP Relay Agent. . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 277
. 277
. 277
. 279
. 282
. 285
. 286
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
v
6.2.2 OS/2 Warp as a DHCP Relay Agent . . . . . . . . . . . . .
6.2.3 DHCP relay considerations . . . . . . . . . . . . . . . . . . . .
6.3 IP masquerading/network address translation . . . . . . . . .
6.3.1 Translation mechanism. . . . . . . . . . . . . . . . . . . . . . .
6.4 Dynamic IP with Network TeleSystems Shadow IPserver .
6.4.1 Installing Shadow IPserver . . . . . . . . . . . . . . . . . . . .
6.4.2 Installation procedure . . . . . . . . . . . . . . . . . . . . . . . .
6.4.3 IPserver configuration overview . . . . . . . . . . . . . . . .
6.4.4 Configuring through IPmanager . . . . . . . . . . . . . . . .
6.4.5 DHCP configuration . . . . . . . . . . . . . . . . . . . . . . . . .
6.4.6 DHCP pools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.4.7 DHCP configurations . . . . . . . . . . . . . . . . . . . . . . . .
6.4.8 Shadow IPserver DNS . . . . . . . . . . . . . . . . . . . . . . .
6.4.9 NetBIOS name services . . . . . . . . . . . . . . . . . . . . . .
6.4.10 Backing up your IPserver configuration. . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 289
. 291
. 291
. 293
. 293
. 294
. 295
. 298
. 299
. 302
. 307
. 309
. 310
. 312
. 314
Chapter 7. Mobile users . . . . . . . . . . . . . . . . . . .
7.1 PPP dial-up . . . . . . . . . . . . . . . . . . . . . . . . . .
7.1.1 Prerequisites . . . . . . . . . . . . . . . . . . . . .
7.1.2 Step-by-step procedure . . . . . . . . . . . . .
7.2 Roaming users . . . . . . . . . . . . . . . . . . . . . . . .
7.2.1 Prerequisites . . . . . . . . . . . . . . . . . . . . .
7.2.2 Automatic connection to file/print domain
7.2.3 Netscape Message of the Day service . .
7.2.4 Custom Netscape INI files . . . . . . . . . . .
7.2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 317
. 317
. 318
. 320
. 322
. 323
. 323
. 334
. 341
. 346
Chapter 8. Security of DHCP and Dynamic DNS . . .
8.1 Security trade-off . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2 RSA public key authentication system . . . . . . . . . .
8.3 Getting more information from the client to DNS. . .
8.4 Presecured domain . . . . . . . . . . . . . . . . . . . . . . . .
8.5 ProxyArec consideration. . . . . . . . . . . . . . . . . . . . .
8.5.1 ProxyArec and option 81 . . . . . . . . . . . . . . . .
8.6 Securing lease allocations . . . . . . . . . . . . . . . . . . .
8.6.1 Preventing access to unauthorized devices . .
8.6.2 Rogue DHCP servers . . . . . . . . . . . . . . . . . . .
8.6.3 Connecting to Untrusted Networks - Firewalls
8.6.4 Connecting through untrusted networks - VPN
8.6.5 TFTP security . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 347
. 347
. 348
. 351
. 353
. 356
. 358
. 360
. 360
. 360
. 361
. 363
. 364
Chapter 9. Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
9.1 Battlefield questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
9.2 Failure events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
vi
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
9.2.1 Severed connections . . . . . . . . . . . .
9.2.2 Facility loss . . . . . . . . . . . . . . . . . . .
9.2.3 Router outages . . . . . . . . . . . . . . . . .
9.2.4 DHCP server problems . . . . . . . . . . .
9.2.5 Name server difficulties . . . . . . . . . .
9.2.6 Other server vulnerabilities . . . . . . . .
9.2.7 Client failures . . . . . . . . . . . . . . . . . .
9.3 Redundancy and fault tolerance in DHCP.
9.3.1 The challenge of fault tolerance . . . .
9.3.2 Cooperative redundancy. . . . . . . . . .
9.3.3 Non-cooperative redundancy . . . . . .
9.3.4 Standby DHCP server arrangement .
9.4 AIX features . . . . . . . . . . . . . . . . . . . . . . .
9.5 Shadow IPserver features . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 366
. 366
. 367
. 367
. 368
. 368
. 369
. 369
. 370
. 370
. 371
. 371
. 380
. 380
Chapter 10. Performance . . . . . . . . . . . . . . .
10.1 Leases . . . . . . . . . . . . . . . . . . . . . . . . . .
10.1.1 Choosing a lease time . . . . . . . . . .
10.1.2 DHCP server differences . . . . . . . .
10.1.3 Multiple leases . . . . . . . . . . . . . . . .
10.2 Monitoring and troubleshooting . . . . . . .
10.2.1 The PING command . . . . . . . . . . . .
10.2.2 The TRACERTE command . . . . . . .
10.2.3 The IPTRACE command. . . . . . . . .
10.2.4 The ARP command . . . . . . . . . . . .
10.2.5 The NETSTAT command . . . . . . . .
10.2.6 The HOST command . . . . . . . . . . .
10.2.7 The NSLOOKUP command . . . . . .
10.2.8 Other utilities . . . . . . . . . . . . . . . . .
10.3 Troubleshooting TCP/IP networks . . . . .
10.3.1 Prerequisites for troubleshooting . .
10.3.2 A bottom-up approach . . . . . . . . . .
10.4 Tuning TCP/IP networks . . . . . . . . . . . . .
10.4.1 An approach to tuning your network
10.4.2 TCP/IP tuning parameters . . . . . . .
10.5 Bandwidth efficiency. . . . . . . . . . . . . . . .
10.5.1 Broadcast traffic . . . . . . . . . . . . . . .
10.5.2 RSVP . . . . . . . . . . . . . . . . . . . . . . .
10.5.3 Communications server . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 381
. 381
. 381
. 382
. 383
. 384
. 384
. 385
. 386
. 387
. 387
. 388
. 389
. 389
. 390
. 390
. 391
. 403
. 403
. 404
. 407
. 408
. 408
. 409
Chapter 11. Managing your network . . . . . . . . . . . . . . . . . . . . . . . . . . 411
11.1 Remote administration of DHCP and DDNS . . . . . . . . . . . . . . . . . . 411
11.1.1 No extra software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
vii
11.1.2 Using a Web server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
Appendix A. Creating DHCP boot diskettes . . . . . . . . . . . . . . . . . . . . . 429
A.1 OS/2 Warp boot diskettes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
A.1.1 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
A.1.2 Step-by-step procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
A.1.3 Notes on the step-by-step procedure . . . . . . . . . . . . . . . . . . . . . . . 438
A.2 DOS boot diskettes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
A.2.1 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
A.2.2 Step-by-step procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
A.2.3 Notes on the step-by-step procedure . . . . . . . . . . . . . . . . . . . . . . . 444
Appendix B. Application issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
B.1 DB2 Universal Database. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
B.2 TME 10 Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
B.3 Netfinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
B.3.1 Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
B.3.2 Client (services) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
B.4 Netscape Navigator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
Appendix C. Using the additional material . . . . . . . . . . . . . . . . . . . . . . 453
C.1 Using the CD-ROM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
C.2 Locating the additional material on the Internet . . . . . . . . . . . . . . . . . . . 454
C.3 Second edition CD-ROM contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454
C.4 First edition CD-ROM contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454
Appendix D. DHCP options (RFC 2132) . . . . . . . . . . . . . . . . . . . . . . . . . 457
D.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457
D.2 DHCP and BOOTP options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
D.2.1 Options 0 and 255: Pad and end . . . . . . . . . . . . . . . . . . . . . . . . . . 458
D.2.2 Option 1: Subnet mask . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
D.2.3 Option 2: Time offset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
D.2.4 Option 3: Router . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
D.2.5 Option 4: Time server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
D.2.6 Option 5: IEN 116 (Old) name server . . . . . . . . . . . . . . . . . . . . . . . 459
D.2.7 Option 6: Domain name server. . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
D.2.8 Option 7: Log server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
D.2.9 Option 8: Cookie server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
D.2.10 Option 9: LPR server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
D.2.11 Option 10: Impress server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
D.2.12 Option 11: Resource location server . . . . . . . . . . . . . . . . . . . . . . 461
D.2.13 Option 12: Host name. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
D.2.14 Option 13: Boot file size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
D.2.15 Option 14: Merit dump file name . . . . . . . . . . . . . . . . . . . . . . . . . 462
viii
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
D.2.16
D.2.17
D.2.18
D.2.19
D.2.20
D.2.21
D.2.22
D.2.23
D.2.24
D.2.25
D.2.26
D.2.27
D.2.28
D.2.29
D.2.30
D.2.31
D.2.32
D.2.33
D.2.34
D.2.35
D.2.36
D.2.37
D.2.38
D.2.39
D.2.40
D.2.41
D.2.42
D.2.43
D.2.44
D.2.45
D.2.46
D.2.47
D.2.48
D.2.49
D.2.50
D.2.51
D.2.52
D.2.53
D.2.54
D.2.55
D.2.56
D.2.57
D.2.58
Option 15: Domain name . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
Option 16: Swap server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
Option 17: Root path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
Option 18: Extensions path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
Option 19: IP forwarding enable/disable. . . . . . . . . . . . . . . . . . . . 463
Option 20: Non-local source routing enable/disable . . . . . . . . . . . 463
Option 21: Policy filter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
Option 22: Maximum datagram reassembly size . . . . . . . . . . . . . 464
Option 23: Default IP time-to-live . . . . . . . . . . . . . . . . . . . . . . . . . 465
Option 24: Path MTU aging timeout . . . . . . . . . . . . . . . . . . . . . . . 465
Option 25: Path MTU plateau table . . . . . . . . . . . . . . . . . . . . . . . 465
Option 26: Interface MTU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
Option 27: All subnets are local . . . . . . . . . . . . . . . . . . . . . . . . . . 466
Option 28: Broadcast address . . . . . . . . . . . . . . . . . . . . . . . . . . . 466
Option 29: Perform mask discovery . . . . . . . . . . . . . . . . . . . . . . . 466
Option 30: Mask supplier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Option 31: Perform router discovery. . . . . . . . . . . . . . . . . . . . . . . 467
Option 32: Router solicitation address . . . . . . . . . . . . . . . . . . . . . 467
Option 33: Static route . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Option 34: Trailer encapsulation. . . . . . . . . . . . . . . . . . . . . . . . . . 468
Option 35: ARP cache timeout . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
Option 36: Ethernet encapsulation . . . . . . . . . . . . . . . . . . . . . . . . 468
Option 37: TCP default time-to-live . . . . . . . . . . . . . . . . . . . . . . . 469
Option 38: TCP keepalive interval . . . . . . . . . . . . . . . . . . . . . . . . 469
Option 39: TCP keepalive garbage . . . . . . . . . . . . . . . . . . . . . . . 469
Option 40: Network information service domain . . . . . . . . . . . . . . 470
Option 41: Network information server . . . . . . . . . . . . . . . . . . . . . 470
Option 42: Network time protocol server . . . . . . . . . . . . . . . . . . . 470
Option 43: Vendor-specific information . . . . . . . . . . . . . . . . . . . . 470
Option 44: NetBIOS over TCP/IP name server option . . . . . . . . . 471
Option 45: NetBIOS over TCP/IP datagram distribution server . . 472
Option 46: NetBIOS over TCP/IP node type . . . . . . . . . . . . . . . . 472
Option 47: NetBIOS over TCP/IP scope. . . . . . . . . . . . . . . . . . . . 472
Option 48: X-Window system font server option . . . . . . . . . . . . . 473
Option 49: X-Window system display manager . . . . . . . . . . . . . . 473
Option 64: Network Information Service (plus) domain . . . . . . . . 473
Option 65: Network Information Service (plus) server . . . . . . . . . 474
Option 68: Mobile IP home agent . . . . . . . . . . . . . . . . . . . . . . . . . 474
Option 69: Simple Mail Transport Protocol (SMTP) server . . . . . 474
Option 70: Post Office Protocol (POP3) server . . . . . . . . . . . . . . 474
Option 71: Network News Transport Protocol (NNTP) server . . . 475
Option 72: Default World Wide Web (WWW) server . . . . . . . . . . 475
Option 73: Default Finger server . . . . . . . . . . . . . . . . . . . . . . . . . 475
ix
D.2.59 Option 74: Default Internet Relay Chat (IRC) server . . . . . . . . . . 476
D.2.60 Option 75: StreetTalk server. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
D.2.61 Option 76: StreetTalk Directory Assistance (STDA) server . . . . . 476
D.3 DHCP (only) options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
D.3.1 Option 50: Requested IP address . . . . . . . . . . . . . . . . . . . . . . . . . 477
D.3.2 Option 51: IP address lease time . . . . . . . . . . . . . . . . . . . . . . . . . . 477
D.3.3 Option 52: Option overload. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477
D.3.4 Option 53: DHCP message type . . . . . . . . . . . . . . . . . . . . . . . . . . 478
D.3.5 Option 54: Server identifier. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 478
D.3.6 Option 55: Parameter request list. . . . . . . . . . . . . . . . . . . . . . . . . . 478
D.3.7 Option 56: Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
D.3.8 Option 57: Maximum DHCP message size . . . . . . . . . . . . . . . . . . 479
D.3.9 Option 58: Renewal (T1) time value . . . . . . . . . . . . . . . . . . . . . . . . 479
D.3.10 Option 59: Rebinding (T2) time value . . . . . . . . . . . . . . . . . . . . . . 480
D.3.11 Option 60: Vendor class identifier. . . . . . . . . . . . . . . . . . . . . . . . . 480
D.3.12 Option 61: Client identifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
D.3.13 Option 66: TFTP server name . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
D.3.14 Option 67: Boot file name . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
D.4 Unofficial DHCP options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
D.5 Options supported by popular operating systems . . . . . . . . . . . . . . . . . 483
D.5.1 Servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
D.5.2 Clients. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Appendix E. Special notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485
Appendix F. Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
F.1 IBM Redbooks publications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
F.2 IBM Redbooks collections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 490
F.3 Other resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 490
F.4 Referenced Web sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
F.5 Useful Web and FTP sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
How to get IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
IBM Redbooks fax order form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496
Glossary and abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
IBM Redbooks evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 517
x
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Preface
Beyond DHCP — Work Your TCP/IP Internetwork with Dynamic IP explores
important network design issues for today’s modern mixed intranets. The
authors discuss many different platforms, including Windows 95, Windows
98, Windows NT, OS/2 Warp, OS/2 Warp Server, AIX, Macintosh, WorkSpace
On-Demand, Linux, Solaris, and others. The book examines how to connect
these systems in a reliable, flexible, high-performance TCP/IP network.
When confronted with thorny TCP/IP problems, network specialists, support
staff, network managers, and other technicians, can refer to this book to
troubleshoot network outages and to prevent them from recurring. The
authors emphasize cost-effective remote management and dynamic setup,
providing real world examples of DHCP and DDNS technologies applied to
roaming users, intelligent Web browsing, software distribution, printing, PPP
dial-up, and other network needs.
This publication helps TCP/IP network operators establish a solid foundation
for future growth without undue expense. Enterprise network staff will
appreciate the sections that explore high end technologies, including Network
TeleSystems’ Shadow IPserver, IBM Communications Server, and UNIX
platforms. All network managers should find the authors’ advice on security
well worth reading.
This second edition updates the contents of the previous redbook to include
OS/2 Warp Server for e-business with TCP/IP Version 4.21, Windows 98 and
Windows 2000, and adds more information about dynamic IP configuration on
AIX. It also takes a fresh look at dynamic IP at the protocol level. Most
sections were verified to ensure currency.
The team that wrote this redbook
This redbook is the second edition of a redbook that was published originally
in August, 1998. This update was produced by a team of specialists from the
U.S. and South Africa working at the International Technical Support
Organization in Austin, Texas.
Oscar Cepeda is an Advisory Software Engineer and Project Leader at the
International Technical Support Organization, Austin Center. He writes
extensively and teaches IBM classes worldwide in areas including OS/2,
OS/2 Warp Server, WorkSpace On-Demand and Windows client networking.
Before joining the ITSO in 1995, Oscar worked in IBM U.S. Availability
Services as an I/T Specialist. He has 11 years of experience with IBM.
© Copyright IBM Corp. 1998, 2000
xi
Bob Chambers is a Network Top Gun with IBM Global Services, and a
Certified Network Expert (see http://www.cnx.org). For the past six years he
has focused on baseline analysis and problem determination on local and
wide area networks. Bob joined IBM in 1982 after receiving a degree from
McNeese State University in Lake Charles, Louisiana. He can be contacted
via e-mail at: [email protected]
Julian Mosca is an Advisory IT Specialist in Service Delivery with IBM Global
Services in South Africa. He has seven years of experience in Network
Administration. Julian has worked at IBM for four years and has recently
completed his B-Tech degree in Information Technology at the Technikon
Witwatersrand. His areas of expertise include electronic software distribution,
client/server support and monitoring, security, and LAN and WAN
administration. He has also worked extensively with the OS/2 and Windows
platforms.
Matt Robbins is an RS/6000 Technical Sales Specialist in Dallas, Texas. He
has over five years of experience with RS/6000 and AIX. His areas of
expertise include UNIX, TCP/IP, and designing e-business solutions for
Internet security and Web traffic. He attended the University of North Texas
as a student of computer science.
xii
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 1. The redbook team: (left to right) upper row: Bob Chambers, Oscar Cepeda; lower row:
Julian Mosca, Matt Robbins
Thanks to the following people for their invaluable contributions to this project:
IBM Corporation
http://www.ibm.com
•
•
•
•
•
•
•
David Babbitt, AIX TCP/IP Development, Austin, Texas
Doug Foulds, DB2 Information Development, Toronto, Canada
Tim Freitas, Networking I/T Specialist, Waltham, Massachusetts
Preetom Goswami, TCP/IP Software Engineer, Bangalore, India
Newton Hammet, AIX SupportLine, Austin, Texas
Andreas Limmer, Software Solutions Development, Boeblingen, Germany
Amol Mahamuni, Customer Support, Austin, Texas
xiii
IBM International Technical Support Organization (ITSO)
http://www.redbooks.ibm.com
• Richard Hawes, ITSO Austin, Texas
• Temi Rose, ITSO Austin, Texas
• Theo Winkelmann, ITSO Austin, Texas
Network Associates
http://www.nai.com
• Glenn Karle
Network Associates supplied our team with Sniffer Basic.
Network TeleSystems (NTS)
http://www.nts.com
• Bob Baumann
• Russell Young
We appreciate NTS’ support for Shadow IPserver in our labs.
We also thank Richard Grey of Brookshire Grocery Company for his input.
This redbook is an update of the original redbook which was written by the
following team:
• Uwe Zimmermann is an Advisory Systems Engineer and Project Leader
formerly at the ITSO Austin Center. He has more than eleven years of
experience in heterogeneous networking environments. Before joining the
ITSO, Uwe worked in an IBM branch office in Stuttgart, Germany, and was
in charge of large accounts as a Networking Systems Engineer. His areas
of expertise include OS/2 LAN/Warp Server, Windows 95/NT, NetWare,
software distribution (CID), systems management, Dynamic IP, and
network computing.
• Peter Degotardi is a LAN Systems Specialist within IBM Global Services
Australia. He has thirteen years of experience in networking with a wide
variety of computing devices and operating systems. His current areas of
expertise include OS/2 Warp, OS/2 LAN/Warp Server, Windows NT,
TCP/IP, Netfinity, REXX, and software distribution via CID.
• Michael McDaniel, a member of ACM, CPSR, and IEEE, and is also
owner of The Fourth Crusade, a consulting firm specializing in automation
and communications. For the last 21 years, he has worked with high
availability systems in the industrial, financial, and telecommunications
fields. Mr. McDaniel’s company is an IBM Premier Business Partner
( http://www.fourthcrusade.com). His areas of expertise include DB2
(various platforms), software distribution (CID, Netfinity), OS/2 Warp
xiv
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Server, Dynamic IP, and integration with UNIX and Windows 95/NT
networks. He holds various certifications from IBM, including DB2 and
OS/2 Warp Server.
• Toshimasa Shimizu is a Senior IT Specialist with IBM, Japan. He was
with ITSO Austin Center for five years supporting OS/2 LAN Server, OS/2
Warp Server, and OS/2 Warp-related products. For more than twenty
years, he has been a systems engineer for large account customer
projects in Japan, providing LAN system designs, distributed systems, and
networking systems.
• Timothy Sipples is Advisory Technical Marketing Specialist with IBM. He
is based in Chicago, Illinois. He specializes in software distribution issues,
Java, and thin clients such as WorkSpace On-Demand. He also has
extensive experience with Lotus Domino, TCP/IP, and the interoperability
of OS/2 Warp with NetWare. Mr. Sipples is coauthor of the IDG/IBM Press
book IBM’s Official OS/2 Warp Frequently Asked Questions.
• Akiko Ueno is an IT Engineer with IBM Japan in the PC Server Group.
She specializes in customer service for OS/2 Warp Server and IBM PC
Servers running Windows NT. She joined IBM three years ago.
Thanks to the following people for their invaluable contributions to this project:
IBM Corporation
http://www.ibm.com
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Greg Althaus, TCP/IP Applications and Kernel Development, Austin, TX
Barry Arndt, LAN Transport Development, Austin, TX
Mel Bryant, OS/2 Warp Server Customer Focal Point, Austin, TX
Esther Burwell, Dynamic IP and Management Development, Raleigh, NC
Charlotte Davis, Dynamic IP and Management Development, Raleigh, NC
Bruce Faulkner, Dynamic IP and Management Development, Raleigh, NC
Steven French, OS/2 Server for e-business, Austin, TX
Pratik Gupta, Dynamic IP and Management Development, Raleigh, NC
Wayne Ha, DB2 Development, Toronto, Canada
Robert Hanner, Network Computing Software, Austin, TX
Bill Hartner, OS/2 Warp Server SMP Development, Austin, TX
Juliana Hsu, DB2 Development, Toronto, Canada
Steven King, OS/2 Warp Server Product Manager, Austin, TX
Saravana Kumar, OS/2 Support, India
Rachele Powell, Dynamic IP and Management Development, Raleigh, NC
Pankaj Sinha, OS/2 Support, India
IBM’s International Technical Support Organization (ITSO)
http://www.redbooks.ibm.com
xv
• Marcela Adan, ITSO Rochester, Minnesota
• Martin Murhammer, ITSO Raleigh, North Carolina
• Elizabeth Barnes, Marcus Brewer, and Tara Campbell, ITSO Austin, Texas
• Steve Gardner, Jasenn McNair, ITSO Austin, Texas
Hewlett-Packard Corporation
http://www.hp.com
• Wylie McDonald, Austin
• Cyndi Watson, Austin
Network Associates
http://www.nai.com
• Robert S. Kusters, Product Manager
• Paul Farr, Director of Product Marketing
Network TeleSystems (NTS)
http://www.nts.com
• Bob Baumann, Dallas
• Russ Young, Austin
Comments welcome
Your comments are important to us!
We want our Redbooks to be as helpful as possible. Please send us your
comments about this or other Redbooks in one of the following ways:
• Fax the evaluation form found in “IBM Redbooks evaluation” on page 517
to the fax number shown on the form.
• Use the online evaluation form found at http://www.redbooks.ibm.com/
• Send your comments in an Internet note to [email protected]
xvi
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 1. TCP/IP overview
Many excellent publications describe TCP/IP and the Internet in a
comprehensive fashion. References to several of these publications, including
requests for comments (RFCs), are included in F.3, “Other resources” on
page 490.
This chapter provides a short introduction to TCP/IP so you can become
acquainted (or reacquainted) with TCP/IP addressing and other basics. What
we have included here are what could be considered pre-requisites to the
TCP/IP theories covered later in this book. If you wish to read a more in-depth
TCP/IP overview, TCP/IP Tutorial and Technical Overview, GG24-3376, is an
excellent source.
After reading this chapter, you should understand what IP addresses,
subnets, routers, and name servers are. You should also learn what the
differences are between Class A, B, C, and D addresses. Static versus
dynamic addressing is discussed, along with basic information on legacy
protocols such as NetBEUI.
If you are already familiar with basic TCP/IP concepts, you may only need to
briefly skim this chapter.
1.1 Network protocols
Computer networks simply deliver bits of information from one point to
another. One requirement for transmitting such information is that the
computer systems on each end speak the same language or protocol.
You are already familiar with many protocols and how they are defined. For
example, to address a regular letter or package in the United States you need
to write the destination address on the front of the envelope. That address
could look like:
John Doe
Your Company
Any Street
City, TX 78758
The protocol for U.S. mail requires a specific location where you write this
destination address (the front of the envelope), the name of the recipient, the
business name (if any), the street address (perhaps with a suite or apartment
number), the city, state, and a postal (zip) code. Additional requirements may
© Copyright IBM Corp. 1998, 2000
1
apply, such as the amount of postage, a return address, proper packaging,
and so on. All these requirements define the U.S. Postal Service protocol.
Computer network protocols require similar information in a precise format. A
package of information sent over the network is called a datagram.
Datagrams usually include at least a destination address, source address
(where it came from), length (size of package), error detection information
(such as a checksum), and package contents (the bits and bytes of
information being carried).
Transmission Control Protocol/Internet Protocol (TCP/IP) is one of the most
popular families of network protocols. It also happens to be the one used as
the basis for the Internet.
Many people think of TCP/IP in terms of layers or levels of functions. At the
lowest layer, the network interface (such as a LAN) carries the network traffic
over wires or other connections. The highest layer, the application layer (such
as a Web browser), uses the various TCP/IP services to communicate. In
between these are two additional layers called the transport and internetwork
layers.
The transport layer facilitates communication between applications, whether
they are on the same or different systems. The main transport layer protocol
is called TCP, and it can determine whether a message has been received or
not at the other end of the connection. An alternative is UDP (User Datagram
Protocol), which simply sends messages without checking to see whether the
system at the other end has received each part. Applications that need
maximum performance and that verify delivery themselves often use UDP.
Internet Protocol (IP) is one of the internetwork layer protocols, and it is
responsible for properly routing datagrams to other computers across the
network. (Other Internet layer protocols include ICMP, IGMP, ARP, and RARP,
discussed in “Other TCP/IP terms” on page 26.) IP depends on several
important addresses in order to keep track of where messages should be
delivered.
1.2 IP addresses
IP uses addresses to specify both the source and destination systems on a
TCP/IP network. Each address consists of 32 bits, usually broken into four
decimal numbers separated by dots (.). Each decimal number represents an
8-bit byte (an octet) in the address.
2
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
For example:
00001001 01000011 00100110 00000001
9
.
67
.
38
.
1
32-bit address
decimal address
Each address can also be separated into two logical parts:
• Network address
The network address is a lot like a postal code, because it identifies which
region (or section) of the total network contains the system.
• System (or machine, or host) address
The system address is similar to an apartment or suite number, because it
specifically identifies a particular system within that region.
As shown in Figure 2, IP addresses belong to one of four classes depending
on how the entire 32-bit address is split (a fifth class, class E, is not
commonly used).
Figure 2. IP Classes
Class A
Class A addresses use 7 bits for the network address portion and
24 bits for the host address. With Class A addresses, there are
126 (27 -2) possible networks (regions) with 16,777,214 (224 -2)
possible hosts in each, for a total of over 2 billion addresses. One
Chapter 1. TCP/IP overview
3
bit is used to identify the address as Class A, to distinguish it from
other classes.
For example, the following Class A address can be broken apart
into its network and host addresses:
00001001 01000011 00100110 00000001
9
.
67
.
38
.
1
^******* ++++++++ ++++++++ ++++++++
The first bit (marked with ^), a zero, identifies this IP address as
Class A. The next 7 bits (*) provide the network number (9). The
remaining bits (+) identify the specific host within network 9, in this
case 67*(216 )+38*(2 8 )+1, or 4,400,641. In other words, this IP
address identifies the 4,400,641th system in the 9th network
region.
Class B
Class B addresses use 14 bits for the network portion and 16 bits
for the host portion. These addresses provide an additional
16,382 (214 -2) networks with 65,534 (216 -2) hosts each, a total of
over one billion additional addresses. The first two bits of a Class
B address are 1 and 0.
Class C
Class C addresses use 21 bits for the network part and 8 bits for
the machine part, providing 2,097,150 (221 -2) networks with 254
(2 8 -2) hosts each, a total of over half a billion addresses. Class C
addresses begin with 110.
As you can see, with Classes A through C and 32-bit addresses,
TCP/IP can provide service for a theoretical maximum of
approximately 3.5 billion different hosts.
Class D
Class D addresses are reserved for multicasting, a limited form of
broadcasting only to other hosts sharing the same Class D
address. Class D addresses begin with bit 1110.
Class E
Class E addresses (beginning with 11110) are not widely used at
this point in time and are reserved for future use.
In addition, some special addresses are reserved and cannot be assigned to
actual systems on the network. These special addresses include:
All bits 0
4
Means this. For example, if the network address part is
set to 0, the host address refers to a system on this (its
own) network. When making initial contact on the
network, a system may use this method if it does not
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
know the network address. Other systems will reply with
the proper network address filled in, and this proper
network address can be recorded for future use.
All bits 1
Means all. For example, if the host address is set to all
ones, the IP address identifies all systems within that
particular network region. In other words, a Class B
address of 128.2.255.255 refers to all systems on
network 128.2. Such an address is also called a directed
broadcast address, because it contains a valid network
address and a broadcast (all ones) host address.
Loopback
The Class A network 127 (including addresses, such as
127.0.0.1) is defined as the loopback network. Systems
will automatically route traffic destined for these
addresses back into the same system without ever
communicating across the real network. Loopback
addresses are often used for testing new software, to
separate network problems from simple programming
errors.
Private addresses
Several addresses have been reserved for private
networks which are not directly connected to the
Internet. These addresses include the Class A group of
addresses in network 10, the 16 Class B groups of
addresses in networks 172.16 through 172.31, and the
256 Class C groups of addresses in networks 192.168.0
through 192.168.255. Web servers, FTP sites, and other
systems available to the public on the Internet will never
have addresses beginning with 10, 172.16 through
172.31, or 192.168. These addresses are quite useful
for testing purposes, or for totally private use, and you
can use them without having to contact someone to
reserve real addresses.
You can determine whether an address is Class A, B, C, D, or E by simply
examining the first octet, as shown in Table 1.
For an address of X.0.0.0, if X is...
Table 1. Class determination table
...from (lowest)...
...to (highest), then...
Class
00000000
0
011111111
127
A
Chapter 1. TCP/IP overview
5
...from (lowest)...
...to (highest), then...
Class
10000000
128
10111111
191
B
11000000
192
11011111
223
C
11100000
224
11101111
239
D
11110000
240
11111111
255
E
Using Table 1 on page 5, and taking into consideration that 127.0.0.0 is
defined as the loopback network, Table 2 summarizes the properties of Class
A, B, and C addresses.
Table 2. Class properties
Class A
Class B
Class C
Lowest Network ID
1.0.0.0
128.0.0.0
192.0.0.0
Highest Network ID
126.0.0.0
191.255.0.0
223.255.255.0
125
16,382
2,097,150
16,777,214
65,534
254
Number of
Networks
Number of Hosts
per Network
1.3 IP subnets
Suppose your company, MegaHuge Industries, is assigned a Class A address
for all its systems around the world. Therefore, you have 16,777,214 possible
IP addresses available to assign to all your systems. However, if you have
many different buildings scattered across the globe, it can be hard to manage
such a large number of addresses. It might be much easier to have
individuals in each location (or even within a particular department) manage
their own smaller sets of addresses. In addition, it's often bad for network
performance to have broadcast traffic throughout the Class A network. To cut
down on network congestion, your company may wish to divide this huge
range of addresses into more manageable chunks.
Subnets were introduced to help solve these problems. By sacrificing the total
number of addresses (16,777,214 for a Class A), you can divide your network
into separately managed subnets (smaller groups of addresses). Class A, B,
6
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
and even C addresses can be divided into subnets. Then, address
assignments within a subnet can be performed locally, without having to
contact a central authority to obtain additional addresses. The whole network
(with subnets) still appears to be one IP network to the outside world.
For example, consider the Class A address 9.67.38.1. The network address is
9 and the host address is 67.38.1. Subnets simply extend this basic address
by treating part of the host address as a subnetwork address. IP addresses
then consist of the class identifier (0 for Class A in this example), the network
address (9), the subnetwork address (for example, bits 8 to 25), and the host
address (for example, the remaining bits 26 to 31).
A bit mask, known as the subnet mask, is used to identify which bits are part
of the subnet address and which are still part of the host address. This 32-bit
subnet mask has the bits for the host address set to 0 and all the other bits
set to 1. By convention, the subnet mask never has a one following a zero.
Both the network address and the subnet address are masked by ones.
However, only the subnet number needs to be masked.
So for 9.67.38.1, with only bits 26 to 31 representing the host address, the
subnet mask would be:
11111111 11111111 11111111 11000000
or 255.255.255.192 in decimal format. To extract the subnet base address
(the network address and the subnet address together), a logical AND is
performed. If both bits in a particular bit position, for both the IP address and
the subnet mask, are set to 1, then the result is also 1. Otherwise, the value is
0. In this case:
00001001
11111111
========
00001001
01000011
11111111
========
01000011
00100110
11111111
========
00100110
00000001 = 9.67.38.1
(Class A address)
11------ = 255.255.255.192 (subnet mask)
========
(logical AND)
00------ = 9.67.38
(subnet base address)
and the remainder is:
-------- -------- -------- --000001 = 1
(host address)
Of course the subnet number (by itself) is:
-------- 01000011 00100110 00------ = 68760
Any subnet number can be chosen with the exception of all zeros (this
subnet) or all ones (all subnets).
Chapter 1. TCP/IP overview
7
One disadvantage to using subnets is that the total number of possible IP
addresses available to you decreases. For example, if you have a Class C
network address of 220.23.5 you then have eight remaining bits you control.
Without subnets, or more precisely, using just one subnet, you have 254
possible host addresses. With two subnets you sacrifice two bits carving out
subnets 01 and 10 (remember all ones and all zeros are reserved) leaving 6
bits remaining. Therefore, you're left with two subnets with up to 62 host
addresses in each for a total of 124 host addresses giving you less than half
the previous number.
Bear in mind both the advantages and disadvantages of using subnets. Using
Table 3, you can easily determine the number of available subnets and hosts
in each subnet. By subnetting your network, you can more easily mix different
network technologies such as Ethernet and token-ring. This can allow you to
overcome limitations to the number of hosts per segment and minimize
network congestion by reducing broadcast traffic.
If you have...
...then your subnet mask is...
255.X.0.0
Class A
Class B
255.255.X.0
Class C
255.255.255.X
...where your choice of X yields...
Table 3. Number of subnets and hosts per subnet (partial)
8
Number of possible hosts in each subnet
Number of
required
subnets
Class A
Class B
Class C
0
1
16,277,214
65,534
254
128
invalid
invalid
invalid
invalid
192
2
4,194,302
16,382
62
224
6
2,097,150
8,190
30
240
14
1,048,574
4,094
14
248
30
524,286
2,046
6
252
62
262,142
1,022
2
254
126
131,070
510
invalid
255
254
65,534
254
invalid
...
...
...
...
invalid
X
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Note
A Class C network cannot contain more than 62 subnets. Also, Table 3 only
shows the possible number of subnets and hosts per subnet when only one
octet is used for subnetting. For Class A and Class B networks, additional
bits can be used for subnetting, as suggested by the last row in the table.
Supernetting, the opposite of subnetting, treats multiple networks as if they
belong to one larger network. For example, a company that may need up to
2,032 possible hosts could be assigned eight Class C network IDs
(8*254=2,032) to form one supernet. For example, if those eight Class C
networks are 220.78.168 through 220.78.175, then the subnet mask to create
the supernet would be 255.255.248.0, which corresponds to the subnet mask
used for a Class B network with 30 subnets of 2,046 possible hosts in each.
Supernets are formed primarily to simplify routing on a TCP/IP network.
1.4 IP routing
As mentioned above, one reason for dividing a network into subnets is to cut
down on the amount of broadcast traffic throughout the network. Generally, a
system on one subnet can send IP datagrams to a system on another subnet
only by working through an intermediary called an IP router or gateway.
Again, let's use the Postal Service as an example. A mail delivery system
without subnets would mean that every delivery truck visits every home and
business until all the packages and letters are delivered. Such a system might
work and pickup service would be quite frequent, but delivery would take a
long time. Consequently, the Postal Service uses subnets by using one truck
serving each particular neighborhood. Packages and letters picked up by that
truck are routed to one or more central handling facilities then further routed
to other trucks on the delivery side. However, if your package or letter is
destined for your neighbor's house (meaning it is on the same subnet) there
is very little routing involved and the same truck will deliver that particular
mail. The size of each truck route, the number of central facilities, the speed
of transfers, and so on determine the overall level of service and
performance.
Similarly, there are two types of IP routing. Direct routing (or direct delivery)
means that an IP datagram can be sent directly to another system which is on
the same subnet without involving an intermediate gateway. Indirect routing
occurs when the destination host is not on a network directly attached to the
source machine. One or more gateways must handle the traffic, and the IP
Chapter 1. TCP/IP overview
9
address of the first gateway must be provided to the machine trying to send
the datagram. This address is called the gateway address or router address.
Each system on the network maintains a routing table to help determine
which type of routing to use, indirect or direct, to reach another system. Three
types of routing can be found in the table:
1. Direct routes
2. Indirect routes
3. Default routes (in case the destination IP network is not found
elsewhere in the table)
A routing table might look like this:
destination
129.7.0.0
128.15.0.0
128.10.0.0
default
127.0.0.1
router
129.7.0.1
128.15.0.1
128.10.0.5
128.10.0.5
127.0.0.1
interface
ethernet
modem
Token-Ring
Token-Ring
loopback
1.5 Assigning IP addresses
In short, there are generally three pieces of information a system needs in
order to start communicating on a TCP/IP network: an IP address to uniquely
identify that system on the network, a subnet mask to help divide that 32-bit
address and determine the subnet and network parts, and at least one default
router address. If a machine does not need to communicate beyond its
immediate subnet, the router address and the machine’s own IP address are
set the same. These three values represent the bare minimum needed for a
system to participate in the TCP/IP world, and they are vital. However, with
networks changing so quickly, manually programming these values into each
and every device attached to the network and reprogramming them as they
change can quickly become tiresome.
IP addresses assigned to systems manually by changing a setting at each
individual system are called static addresses. BOOTP and DHCP can be
used to assign dynamic addresses.
1.5.1 BOOTP and the Dynamic Host Configuration Protocol (DHCP)
BOOTP, the BOOTstrap Protocol, was one of the first attempts to automate
delivery of these critical values. A client system such as a PC can boot up
and obtain the address and IP configuration needed to connect to the network
from a BOOTP server where all the information is kept and managed
centrally.
10
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Although BOOTP is still widely used, its successor, Dynamic Host
Configuration Protocol (DHCP), provides much more flexibility. IP
configurations can be used and reused according to rules set in the DHCP
server without having to make modifications for every change in the network.
Also, many DHCP servers can still provide IP configurations to older BOOTP
clients. A well designed DHCP server can save a lot of work and help get your
TCP/IP network up and running more quickly. If you'd like to start exploring
DHCP, read Chapter 2, “Up and Running with DHCP” on page 29, and
Chapter 3, “DHCP server and client interaction” on page 113.
1.5.1.1 Mobility
Why would one wish their clients to have dynamic IP configurations on their
TCP/IP network? DHCP is an answer to the problem of growing
administrative overhead with client configuration for TCP/IP networks. Take,
for example, a user who travels from one office to another across the country.
While their IP configuration may be completely valid in their Dallas office, this
same IP configuration is not likely to work in their Atlanta office. In this event,
the user would not be able to get on the network, and in order to get on would
have to communicate with their local network administrator to obtain a full IP
configuration (IP address, default IP gateway, name server, etc.) in order to
reconfigure their machine for proper network communication. Furthermore,
the user will have to reconfigure their machine with their original host
configuration when they return to Dallas. However, if the user is configured for
DHCP instead of a static IP configuration, and both sites are configured with
DHCP servers, there will be no administrative overhead associated with the
user travelling between Atlanta and Dallas. The user will be required to simply
attach their machine physically to the network and, upon boot, it will get a
dynamic IP configuration from a DHCP server in order to be able to
communicate properly on the local network.
1.5.1.2 Managing a pool of IP addresses
Another example of why one would want to use DHCP is if a company is
running out of available IP addresses. Let us say a company was only able to
obtain 80 addresses for the 100 host machines within their network. They
know at any given time at least 20% of the machines on the network will not
be active due to employees being out of the office for sick days, vacations,
business trips, or the machines simply not being in use. They can get by on
those 80 addresses by using DHCP to manage their available pool of IP
addresses by only assigning an address to a machine that is active on the
network.
Obviously, this scenario will only work properly if users either power off their
machines while not in use or release their DHCP configuration. For a detailed
Chapter 1. TCP/IP overview
11
discussion of how client DHCP configurations are released, please refer to
3.4.7, “DHCPRELEASE” on page 128.
1.6 Domain Name Services (DNS)
Note
Domain Name Services or Domain Name System? While doing
research we found many RFCs (request for comments) and
BCPs (best common practice) that support both terms for DNS,
so you will find us using both Domain Name Services and
Domain Name System throughout this book to describe the
same concept.
DNS servers handle the translation of IP addresses to alpha-numeric host
names to facilitate easy comprehension of an IP topology. There is nothing
wrong with using only addresses to communicate on a network and to keep
track of where everything is within a network, but it is much easier to
remember where merlot1 is in relation to mustard7 than it is to remember
where 192.168.6.2 is in relation to 192.168.10.6. In this scenario merlot1 is
the host name and 192.168.6.2 is the IP address.
Furthermore, DNS servers are directly responsible for translating URLs like
www.ibm.com into their IP addresses like 9.19.24.10, and vice versa. As you
can see, www.ibm.com is much easier to remember and more meaningful than
9.19.24.10. In this scenario, www.ibm.com is the fully qualified domain name
and 9.19.24.10 is the IP address.
Most systems can accept up to three name server addresses, a primary and
up to two backups, contacting each in turn in order to look up the IP address
that corresponds to a particular name. Yet, there are thousands of DNS
servers connected to the Internet, each handling a small portion of the vast
list of names and addresses. To handle these lookups most efficiently, one
DNS server can forward a lookup request it cannot satisfy to another DNS
server. Often, lookup responsibilities are divided according to the dots (.) in
the fully qualified name. For example, charlie may be given one name server
address and that local name server (DNS 1) knows the names and addresses
of all the systems in the ibmus2.ibm.com group. When charlie requests an IP
address for gadget.att.com from DNS 1, DNS 1 may forward the request to
DNS 2 (serving ibm.com), which then forwards the request to DNS 3 (serving
com), which then forwards the request to DNS 4 (serving att.com), which then
answers with the proper IP address for gadget.att.com. In practice, name
servers record many of the most recent lookups to help cut down on network
12
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
traffic. For example, if samuel.ibmus2.ibm.com requests the IP address for
gadget.att.com from DNS 1 just after charlie does, DNS 1 may be able to
provide the answer without bothering DNS 2, DNS 3, or DNS 4.
Traditionally this is setup as a static DNS, which requires the network
administrator to manually edit and update the lookup table whenever an IP
address gets assigned or reassigned to a particular name, or when a name is
no longer used because the system is out of service. Needless to say, this
can have quite a large administrative overhead. We discuss how this process
can be handled dynamically, and therefore with much less administrative
intervention, in 1.6.4, “Dynamic Domain Name Services (DDNS)” on page 22.
1.6.1 Domain names
Names assigned to subnets, networks, or any arbitrary collection of IP
addresses, are called domain names. Therefore, a system's fully qualified
name might be charlie.ibmus2.ibm.com or alice.kingscollege.ac.uk. Domains
are designed to save typing and to provide some structure to the naming of
systems. Generally, you do not have to include the domain name if you are
simply trying to reach another system within the same domain. Subnets and
network addresses can also be assigned names such as com, ibm.net, or
kingscollege.ac.uk. The system that has the job of keeping track of which
names correspond to which IP addresses is called a domain name services
(DNS) server. A DNS server simply contains a table of addresses with their
corresponding names. Lookups can be performed in either direction but
usually a DNS server translates names into numeric addresses on behalf of
clients. For any machine to take advantage of domain names, it must know
the IP address of at least one name server. This address is the fourth critical
piece of information most client machines need in order to get basic TCP/IP
service, although it is optional since it's still possible (albeit cumbersome) to
reach other systems using nothing but numeric addresses.
Note that domain names are not case sensitive. For example,
charlie.ibmus2.ibm.com, CHARLIE.IBMUS2.IBM.COM, and
ChArLiE.ibmus2.IBM.cOm will all be treated by a DNS the same way, and the
DNS will return an IP address of 9.67.38.1 for each of these variations. It's
also quite common to have multiple listings, known as alias names, with
several different names corresponding to the same numeric IP address. An
example of this is that many Web-based companies try to gain a marketing
advantage by registering multiple names, hoping that someone will stumble
into their Web site by typing a generic name.
The domain name space is tree structured. The top-level domains organize
the Internet domain name space. Examples of top-level domains are:
Chapter 1. TCP/IP overview
13
• com: Commercial organizations, such as IBM (ibm.com) and mycompany
(mycompany.com). ibm is a subdomain of the top-level domain com.
• edu: Educational organizations, such as school.edu or university.edu
• gov: Government organizations, such as the National Science Foundation
(nsf.gov)
• uk,fr,jp: Country specific domains for United Kingdom, France, and Japan,
among many others
The tree is limited to 127 levels; this is a limit on subdomains, although there
is no limit on the number of branches at each node.
Each node in the tree is labeled with a name (see Figure 3). The root has a
null label (" "). The full domain name of any node in the tree is the sequence
of names on the path from the node up to the root with a dot between node
names. For example, in Figure 3, if you follow the arrows from the bottom
label to the top from the host: www to the root label you can form the full
domain name for that host: www.as400.ibm.com.
Managed by the
Network Information Center
com
jp
gov
InterNIC
" "
mil
edu
org
fr
uk
university.edu
ibm.com
www.university.edu
as400.ibm.com
www.as400.ibm.com
unive
in
rsity.edu subdo ma
Figure 3. DNS name space
In DNS, each domain can be administered by a different organization. Each
organization can then break its domains into a number of subdomains and
dole out the responsibility for those domains to other organizations. This is
because DNS uses a distributed database where you can manage your own
domain (company.com), or parts of the name space (subdomains) can be
delegated to other servers (department.company.com).
14
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The DNS servers responsible for the top level Internet domains, such as com,
are also called Internet root servers, and they manage information about the
top-level domains. For example, the Internet's Network Information Center
runs the edu domain, but assigns individual schools authority over the .edu
subdomains.
Domains can contain both hosts and other domains (their subdomains). For
example, the ibm.com domain contains hosts such as www.ibm.com, but it
also contains subdomains such as as400.ibm.com.
Domain names are used as indexes into the DNS database.
Each host on a network has a domain name with a DNS server that points to
information about the host. This information may include an IP address,
information about mail routing, and so on.
Why this complicated structure? It is to solve the problems that a host table
has. For example, making names hierarchical eliminates the problem of name
collisions. Domains are assigned unique domain names giving organizations
the freedom to choose names within their domains. Whatever name they
choose it does not conflict with other domain names since it has its own
unique domain name.
For example, we can have several hosts named www, such as www.ibm.com
and www.Acompany.com because they are in different domains managed by
different organizations. See Figure 4 on page 15.
w w w .a s 4 0 0.ibm .c o m
In t erN IC
" "
ib m .co m no d e
www
as 4 00
ibm
co m
go v
m il
edu
o rg
fr
uk
jp
w w w .ib m .c o m
i bm
.c om d om a in
A c o m p an y
w w w .A c o m p a ny .c om
c o m d o m a in
Figure 4. Hosts with the same names in different domains
We can have a host in the same domain that also has the same hostname,
such as www.ibm.com and www.as400.ibm.com because they belong to
different subdomains.
Chapter 1. TCP/IP overview
15
Registration of IP addresses and domain names on the Internet is currently
managed by a central administrative body called InterNIC, the Internet
Network Information Center. InterNIC’s Internet Web page can be found at:
http://www.networksolutions.com/
1.6.2 Domain versus zone of authority
The concept of domains versus zones of authority can be a confusing one.
We do our best to explain it in this section.
One of the main goals of the design of the Domain Name System is
decentralization. This is achieved through delegation. The central DNS
administrator in your company administering the company’s domain can
divide it into subdomains. Each subdomain can be delegated to other
administrators. This means that the administrator delegated to becomes
responsible for maintaining the subdomain.
Domain versus Subdomain
A domain is a subset or subtree of the name space tree. A subdomain is a
subset of the domain.
Figure 5 on page 18 shows the domain mycompany.com as a subset of the
.com name space. Under mycompany.com, there are other subdomains, such
as:
• endicott.mycompany.com
• rochester.mycompany.com
• otherdomain.mycompany.com
Name servers are programs running on a system, such as OS/2 Warp Server,
AS/400, or AIX, with DNS support. In Figure 5 on page 18, the following hosts
are running name server programs:
• as1.mycompany.com
• rst.rochester.mycompany.com
• otherhost.otherdomain.mycompany.com
They are called Domain Name System (DNS) servers, or simply name
servers.
Name servers have information about the part of the domain name space
called a zone or zone of authority. Both domains and zones are subsets of the
domain name space. A zone contains host information and the same data
that the domain contains, excluding the information that is delegated
16
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
somewhere else. If a subdomain of a domain is not delegated, the zone
contains host information and data for the subdomain.
Name servers have complete host information and data for a specific zone.
Name servers are said to be authoritative for the zone for which they have
this complete host information and data.
As shown in Figure 5 on page 18, the mycompany.com domain is divided into
the following subdomains:
• endicott.mycompany.com
• rochester.mycompany.com
• otherdomain.mycompany.com
The zone mycompany.com contains the following hosts:
•
•
•
•
as1.mycompany.com
as2.mycompany.com
as5.mycompany.com
NTserver1.mycompany.com
It also contains the host information and data in the subdomain
endicott.mycompany.com:
• host1.endicott.mycompany.com
• host2.endicott.mycompany.com
The subdomain endicott.mycompany.com has not been delegated, and its
host information and data remain in the mycompany.com zone. The
administration of the endicott.mycompany.com is the responsibility of the
mycompany.com administrator. as1.mycompany.com is the name server that
has complete host information and data for the mycompany.com zone of
authority.
The zone mycompany.com does not contain information in the subdomains
that have been delegated.
A subdomain of mycompany.com is rochester.mycompany.com, and its
administration has been delegated. The zone rochester.mycompany.com
includes host information and data in the following subdomain:
• rochester.mycompany.com
Which includes host information and data in these subdomains:
• rst.rochester.mycompany.com
• host1.rochester.mycompany.com
• host2.rochester.mycompany.com
Chapter 1. TCP/IP overview
17
The DNS server that has complete host information and data for the
rochester.mycompany.com zone is rst.rochester.mycompany.com.
A subdomain of mycompany.com is otherdomain.mycompany.com, and its
administration has been delegated. The zone otherdomain.mycompany.com
includes host information and data in the subdomain
otherdomain.mycompany.com:
• otherhost.otherdomain.mycompany.com
• otherprinter.otherdomain.mycompany.com
• otherserver.otherdomain.mycompany.com
The DNS server that has complete host information and data for the
otherdomain.mycompany.com zone is
otherhost.otherdomain.mycompany.com.
Figure 5. Domain, subdomain, delegation, and zone of authority
1.6.3 Differentiating name servers
Name servers are devices that store the information about the domain name
space. Usually, they have complete information about some part of the name
space or zone. There are several types of name servers, and they are
illustrated in the following sections.
18
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
1.6.3.1 Primary name servers
This is the server that the hosts in the zone of authority are configured on. It
is the server that the DNS administrator configures and maintains. When this
server gives responses to queries from its primary domain files, the
responses are called authoritative. A name server for a primary domain reads
the primary domain configuration information directly from files configured by
DNS administrator and/or updated by dynamic clients.
1.6.3.2 Secondary name servers
This server has the same information as the primary name server. However,
instead of getting its information directly from the DNS administrator
configuring it, it gets its information from another name server through zone
transfers over the network. Secondary name servers are authoritative
servers, just like the primary name servers, and they act just like the
primaries in terms of data storage and data queries.
Primary vs. Secondary Name Server
A DNS server can be a primary name server for one or more domains, as
well as a secondary name server for one or more domains. The terms
primary name server and secondary name server are somewhat
misnomers, because any given name server may be primary from some
zones and secondary for other zones.
A zone transfer is a TCP/IP transfer of domain files from another DNS server
(called a master name server). This is done automatically when the
secondary name server starts and also when the secondary name server
detects that its domain files are downlevel from the master name server’s
domain files. The zone transfer is initiated from the secondary name server.
The zone transfer cannot take place if the master name server is not active.
A secondary name server is used for two reasons: spreading the DNS query
workload over more than one server and as a backup in case the primary
name server stops responding.
The name servers configured at the client probably have no direct
relationship to the zone being queried, and even if they do, the name servers
might all be secondaries, or the first defined name server might be a
secondary. Don’t be confused with the terms primary and secondary name
server at the client which applies when a client is configured with more than
one DNS server for name resolving. If the first name server (also called the
primary name server) does not respond, the client can query the second
name server (also called the secondary name server). When the secondary
Chapter 1. TCP/IP overview
19
name server gives out a response to a query, the response is also called
authoritative. In other words, an answer from a secondary name server is
considered to be just as good as if the answer came from a primary name
server. The same applies if there is a third name server (also called tertiary
name server) defined at the client resolver configuration.
1.6.3.3 Master name servers
This is the name server from which a secondary name server gets its zone
transfer. A master name server can either be a primary name server or
another secondary name server.
1.6.3.4 Caching-only name servers
A name server that does not have authority over any zone is called a
caching-only name server. It gets all of its information by querying. A
caching-only name server’s responses are always non-authoritative.
1.6.3.5 Authoritative name servers
A server that is considered to be authoritative for a domain is either the
primary server or a secondary server for that domain. If another name server
or a client queries either the primary or the secondary name server for
information that they are authoritative for, the response is considered to be
authoritative. Can a name server that is not authoritative over a domain give a
response to a client about that domain and have that response considered an
authoritative response? The answer is yes. If the non-authoritative server
does not know the answer and queries an authoritative name server on behalf
of the client and then returns the answer to the client, this response is
considered to be authoritative. The non-authoritative name server caches this
information. If a second client requests this same information from the
non-authoritative name server (and this information is still in its cache), the
name server gives the response to the client, but now, this same information
is labeled non-authoritative. Why? Because the information in the response
this second time came out of the name server’s cache. Another way of saying
this is that at some point, a non-authoritative response came out of a name
server’s cache.
1.6.3.6 Parent and child name servers
The concept of parent and child domains is equivalent to the concept of
domain and subdomain. Once your domain grows to a certain size, you may
need to distribute management by delegating authority of part of your domain
to one or multiple subdomains. The upper-level domain is the parent and its
subdomains are the children.
20
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The name server authoritative for the parent domain is the parent name
server, and the one authoritative for the subdomain is the child name server.
For example, in Figure 5 on page 18, otherdomain is a subdomain of
mycompany.com. If a DNS server, AS1, is configured to be responsible for the
mycompany.com zone of authority, and the authority for the zone
otherdomain.mycompany.com is delegated to another DNS server, otherhost,
then AS1 is considered to be the parent name server and otherhost is
considered to be the child name server.
1.6.3.7 Root name servers
Internet root name servers know where name servers that are authoritative
for the top-level domains are, and most of the Internet root name servers are
authoritative for the top-level organizational domains (.com, .edu, .net, and so
on). The top-level domain servers have information about the second-level
domain that a given domain is in.
A company can implement internal root name servers. In this case, given a
query for a company’s subdomain, the internal root name server can provide
information for the second-level subdomain the queried subdomain is in.
A root name server is configured in a lower level name server to help it to
navigate the name space tree top down, when it cannot answer a query with
authoritative data or data in its cache.
If we use the example discussed in the previous section, the DNS server
otherhost is authoritative for the zone otherdomain.mycompany.com, shown
in Figure 5 on page 18. The AS1 name server is authoritative for the
mycompany.com zone of authority and is configured as the internal root for
the whole company’s name space. The internal roots can run on host
systems all by themselves, or a given host can perform double duty as an
internal root and as an authoritative name server for other zones. If otherhost
cannot answer a query, it asks its root name server, which is AS1, the DNS
server at the top of the internal name space tree. We stress internal, because
in this example, these DNS servers are only part of an internal network. We
are assuming that the network does not have Internet access; thus, the
Internet com node is not part of this DNS name space tree. Therefore, the
DNS server AS1 in the domain mycompany.com is at the top of the tree. A
root name server can be thought of as the name server at the top of the DNS
name space tree. Remember that the DNS name space tree may be different,
depending on whether the network is an internal network, or if the network
includes the Internet DNS name space.
Chapter 1. TCP/IP overview
21
1.6.3.8 Forwarders
A DNS server can be configured to send queries to which it does not know
the answer to a DNS server called a forwarder name server. Whereas going
to a root name server for help in answering a query can be thought of as
going to the top of the DNS name space tree, going to a forwarder can be
thought of as going sideways in the DNS name space tree for help. The DNS
administrator configures which DNS server is the forwarder. Usually, several
DNS servers are configured to have the same forwarder. Then, the forwarder
name server is configured with the root name servers (for example, the
Internet root name servers). If the forwarders cannot answer the query, they
query the root name servers, get the answer, and cache it. This way, a
forwarder name server can build up a large cache of information. As the
cache increases, chances are that the forwarder will receive a query for which
it has a cached answer. This, in turn, reduces the number of times a root
name server needs to be queried. Using a forwarder name server is an
opportunity to build a large cache of information on one (or just a few) name
servers.
1.6.3.9 Firewall name servers
A firewall name server is a special application of forwarding name servers.
This could be useful when you want to connect a private network (intranet) to
the Internet. You usually do not want all of your private DNS server
information available to the Internet. You provide a more limited content DNS
server to the public Internet (your Firewall DNS server). When a host on your
private network requests an address, it first queries the private DNS server. If
that name server does not have the information desired, it sends the request
to the firewall name server, which forwards the request out to the public
Internet.
Please refer to the Chapter 8, “Security of DHCP and Dynamic DNS” on page
347, to read how to connect your private network (intranet) to an untrusted
network. This could be a customer’s intranet or the global Internet. In either
case, what we show you will help protect your internal resources from the
malicious or just plain curious folks out there.
1.6.4 Dynamic Domain Name Services (DDNS)
Dynamic domain name services allow the dynamic updating of the lookup
table in order to resolve a dynamic or static host name to a dynamic IP
address. Two resource records, the PTR record and the A record, need to be
updated during DHCP client configuration for DDNS to take place. For an
explanation of resource records, see 1.6.5, “Resource record types” on page
23. DHCP servers generally handle the responsibility of updating PTR
22
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
records, and for the updating of A records, the DHCP client must be able to
update the DDNS server with its A record. This capability is generally written
into the DHCP client code, but sometimes additional software is required.
1.6.4.1 ProxyArec
In the event the DHCP server is handling a legacy DHCP client or BOOTP
client that cannot function as a DDNS client, ProxyArec (the A record update
by proxy) must be utilized. When a DHCP server is configured for ProxyArec,
it takes over the responsibility of updating the DDNS server with both the
client’s A and PTR records.
Furthermore, ProxyArec is required in certain situations, such as setting up
dial-in DHCP servers where the clients will be remote and will not have
access to a DDNS server inside the network. There are considerations for
your network if you are mixing dial-in DHCP clients with ProxyArec and local
network DHCP/DDNS clients. This is covered in 8.5, “ProxyArec
consideration” on page 356.
1.6.5 Resource record types
Name server records, called resource records (RR), are the variables for DNS
that keep track of the resolution of names and IP addresses and are divided
into classes for different kinds of networks. We will only talk about the Internet
class of records here. The IN that you see as the second field in most records
means Internet. This list is not complete. See RFC 1035 for all RR types.
Figure 99 on page 141, Figure 100 on page 142, Figure 127 on page 185 and
Figure 128 on page 185, show examples of files that contain each record, so
you can see the format of each one. These records are:
• SOA records are start of authority records, indicating that this name server
is the main source of authoritative information for this domain. There is
only one SOA record in the configuration files, and it is required to be in
the zone and the address files.
• NS records are name server records. Each NS record should indicate a
valid name server. Sometimes, this is not the case when you receive
information from other name servers. As an aside, if you find a name
server you know provides bad information (or no information), you can use
the bogusns directive in your boot file. For example, bogusns 192.168.6.11
would tell your name server not to query this name server. (Hopefully,
there are other name servers that will be able to provide valid information
for you.)
• MX records are mail exchanger records. An MX record indicates a host
machine that will either deliver mail to the addressee, or forward (using
Chapter 1. TCP/IP overview
23
SMTP) the mail to another host that is closer to the addressee. The other
host will either deliver or forward, as necessary. MX records have a third
parameter to prevent mail loops, the preference value. The preference
value is a relative value for each mail host. The lowest value would
indicate the highest preference for delivery (that is, a mailer should
forward undeliverable mail to that host first). A preference value can be
between 0 and 65535 (a 16 bit unsigned value).
• A records are address records. This is a record in the [D]DNS that maps a
hostname to an IP address so that you can resolve an IP address using by
issuing a name. See also PTR records, below.
• PTR are pointer records. This is a record in the [D]DNS database that
maps an IP address to a hostname so that you can resolve a hostname
using its IP address. See also A records, above.
• KEY entries in the DDNS database contain all hostname/domain
name/primary name server settings that have been created and the
associated keys to be able to update those hostnames.
• RP records are the responsible persons for the domain. If you found a
name server that was delivering bad (or no) information, you would
probably want to try to contact the RP for that name server and rectify the
problem prior to adding a bogusns record to your name server.
• TXT records in the DDNS database may provide more information about
how to contact the RP. A text record can contain arbitrary information in
quotes (for example, a telephone number and name). It basically maps
any kind of information to a hostname. This is a feature that was created
as a tool for the system administrator to map a hostname and/or IP
address lease to a real person. This is done by setting labels in the DHCP
server in option 192 (supported by TCP/IP Version 4.1 for OS/2 DHCP
server).
• RR entries are resource records in the nameserver database. A resource
record maps information to a hostname.
• CNAME records map a host alias name to the canonical name of the host
(that is, the fully qualified domain name of the host that is defined by an
address record).
• HINFO records contain host information and are generally not
recommended for use since you do not want to provide information that
might be useful for any potential hackers. However, TCP/IP Version 4.1 for
OS/2 makes use of HINFO for ProxyArec clients. The ProxyArec client’s
MAC address is stored here in an encoded fashion to ensure that other
clients cannot take over other clients’ hostnames.
24
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
1.7 Network security and the Internet
Firewalls help protect an intranet (private TCP/IP network) against unwanted
access from the Internet, such as hackers trying to manipulate a company’s
Web page, industrial saboteurs trying to steal confidential information, or
people who are just plain curious but have no business poking around your
internal network. At the same time, people on the Internet can still get access
to Internet systems. Whereas a router will pass IP packets from one area of a
network to another without scrutiny, a firewall router inspects packets, blocks
any network packets that are suspect, and logs all transmission of packets
from the intranet to the Internet, or vice versa. This process allows internal
users to access information and services on the Internet while also keeping
out all unfriendly or unwanted network traffic coming in from the Internet.
I n d u s t r ia l
S a b o te u rs
In te rn e t
C N N .c o m
C u rio u s
H a c k e rs
F i r e w
a l l
IB M
In te rn a l
N e tw o rk
Figure 6. Firewalls and your internal network
Firewalls vary in capabilities, but the two basic types are proxy and SOCKS
firewalls. Both pass packets from intranet users to the Internet with the
firewall’s address instead of the user’s, fetching data from the Internet by
proxy. This adds an extra layer of protection by keeping the user’s IP address
secret from the Internet, and for a hacker to break into a user’s computer,
they must first know where to attack.
Firewalls can also help link two separate intranets via the Internet (and still
provide security) using a technology called Virtual Private Networks (VPN).
VPNs are set up by creating an IP Tunnel between two intranet firewalls over
the Internet. An IP Tunnel is a network connection where all data sent
between two IP addresses is encrypted. By encrypting all data that flows
Chapter 1. TCP/IP overview
25
between the two network firewalls, all data flowing between any two points on
the separate networks will be encrypted and therefore protected. Chapter 8,
“Security of DHCP and Dynamic DNS” on page 347, takes a look at firewalls
in the context of overall TCP/IP network security.
1.8 IP Version 6 (IPv6)
IPv4 with 32-bit addresses allows us slightly over four billion IP addresses,
which directly translates to only four billion computers allowed on the Internet.
Currently, over two billion addresses are registered to various companies and
individuals and it is projected that within the next five or more years the entire
pool of four billion IP addresses will be exhausted.
IPv6 (Internet Protocol Version 6) is proposed to help alleviate some of the
address constraints and other shortcomings of TCP/IP and IPv4 as more and
more people and devices get onto the Internet. IPv6 will extends today's
TCP/IP environment, so the skills you develop with this book will still be
applicable when these protocol enhancements become available. See RFCs
1883 through 1887 for more information on IPv6.
1.9 Applications that use TCP/IP
Programmers have written many applications that communicate using
TCP/IP. Web browsers (HTTP), FTP (file transfer protocol), Telnet (terminal
emulation), LPR/LPD (printing over TCP/IP), REXEC and RSH (starting
programs remotely over the network), POP/SMTP/IMAP (e-mail services),
NFS (network file sharing), X-Windows (graphics terminal emulation), NNTP
(news), SNMP (network management), and Java, among other examples, can
all use these common services. Although many variations exist for specific
platforms (such as Winsock for Microsoft Windows), a sockets interface
allows programmers to write applications that communicate with TCP/IP.
1.10 Other TCP/IP terms
To work properly, IP protocol needs some additional help to manage traffic
flow efficiently. These terms are explained briefly; please consult the
appropriate references for more detailed information:
• Internet Control Message Protocol (ICMP), a part of IP, helps report errors
in datagram delivery. ICMP can also help discover routers and maximum
transmission units (MTUs) (see below) along the path the datagram
travels. PING, the popular TCP/IP application used to check the
26
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
connection between two systems on the network, uses ICMP. RFC 792
describes ICMP in detail.
• The maximum transmission unit (MTU) is the size of the IP datagram,
which can be adjusted depending on network conditions. All systems on a
TCP/IP network are required to handle MTUs ranging from 576 to 65,535
octets. Each datagram typically packages 20 octets of identifying
information, such as the destination address, as part of the structure.
• Internet Group Management Protocol (IGMP) allows systems to
participate in IP multicasts and to cancel such participation. Additionally,
IGMP provides routers with the ability to check hosts to see if they are
interested in participating in multicasts. RFCs 966, 1112, and 1458
discuss IP multicasting.
• Address Resolution Protocol (ARP) maps IP addresses to hardware
addresses on a network. (Hardware addresses are often Ethernet or
token-ring network adapter addresses consisting of 12 hexadecimal
digits.) Reverse ARP (RARP) provides the reverse lookup. See RFC 826
for more information on ARP.
• SLIP (RFC 1055) and PPP (RFCs 1717 and 1661) provide TCP/IP
services over serial lines, such as modems and ISDN (digital telephone)
connections. Chapter 7, “Mobile users” on page 317, provides much more
information on PPP.
• Traditional LAN (Local Area Network) protocols can be carried by TCP/IP,
to provide file and print sharing services for PCs without contaminating a
TCP/IP network with other protocols. (Many computer networks are
perfectly capable of mixing protocols, but many network managers prefer
to standardize on one protocol.) For example, RFC 1234 describes how
Novell’s IPX operates over TCP/IP; so traditional NetWare servers and
clients can communicate via TCP/IP. NetBIOS over TCP/IP (used by the
IBM OS/2 Warp Server and clients, IBM DOS/Windows LAN Services,
Microsoft Windows 95, Windows NT, and Windows for Workgroups), and
explained by RFCs 1001 and 1002, supports applications written to the
NetBIOS programming interface. NetBIOS isn't actually a protocol. Like
the sockets interface, it's a widely accepted way of writing network-savvy
applications. The NetBIOS interface can be supported by a number of
underlying network protocols, including NetBEUI (NetBIOS over IEEE
802.2), TCP/IP (TCPBEUI), and even SNA. Like TCP/IP, NetBIOS also
relies on names to communicate with other systems on the network,
although these names behave much differently. A NetBIOS Name Server
(NBNS) can help provide better NetBIOS service over a TCP/IP network.
Consult Chapter 5, “Integrating file and print services” on page 227 for
details on NetBIOS over TCP/IP issues.
Chapter 1. TCP/IP overview
27
28
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 2. Up and Running with DHCP
In this chapter, we show how to set up a simple TCP/IP network with one
subnet using dynamic IP. Each system connected to this simple network will
be provided an address, subnet mask, router address, and name server
information to permit the most basic connection. These critical network
connection values will be provided using DHCP, and we will explain why
having dynamic addressing helps your network run more smoothly.
Furthermore, Chapter 3, “DHCP server and client interaction” on page 113,
explains in detail all interactions between DHCP clients and DHCP servers.
DHCPOffer
DHCPDiscover
Client Sends
Server Sends
1
2
DHCP Cycle
End
3
Client Sends
DHCPRequest
Server Sends
DHCPAck
Figure 7. DHCP cycle
Here you will discover how to connect just about any type of server,
workstation or printer to your DHCP network for true interoperability, as
Figure 7 indicates. As the title suggests, this chapter should get you up and
running quickly. The network you create in this chapter can be useful in many
small environments. In later chapters, we demonstrate how to build upon the
basics introduced here.
© Copyright IBM Corp. 1998, 2000
29
2.1 Why DHCP?
Computer networks always seem to be changing. New devices (PCs, printers,
and so on) must be attached, old devices disconnected, new branches
added, mobile workers hosted, temporary employees accommodated, all
these changes can happen every day. Managing all that change can prove a
major undertaking without systems that respond automatically to changing
demands.
On a TCP/IP network, each system must have an IP address, subnet mask,
and router address (at a minimum) in order to communicate, as discussed in
Chapter 1, “TCP/IP overview” on page 1. At least one name server address
should generally be provided as well. Unless your network contains only a
small number of systems, which rarely (if ever) change, managing all these
address settings manually can become overwhelming. Using a DHCP server
to manage these addresses for you can help make your TCP/IP network more
reliable, flexible, and less time consuming.
2.2 OS/2 Warp Server for e-business as a DHCP server
OS/2 Warp Server for e-business makes a reliable DHCP server, with
sophisticated management features and excellent performance.
This section deals mainly with OS/2 Warp Server for e-business, but if we
make mention of OS/2 Warp Server Version 4, then we are referring to OS/2
Warp Server Version 4 Entry, Advanced or SMP.
The IBM TCP/IP Version 4.21 for OS/2, which provides the DHCP capabilities
of OS/2 Warp Server for e-business, is part of the original package, and may
be installed on OS/2 Warp Server Version 4. TCP/IP Version 4.1 is available
for OS/2 Warp Server Version 4, and may be downloaded from IBM Software
Choice at the following Internet Web site:
http://www.ibm.com/software/os/warp/swchoice/
Note: Some features in the IBM Software Choice catalog, such as TCP/IP
Version 4.1 for OS/2, are fee-based features.
2.2.1 Installation
Be sure to select the DHCP server function during the installation of OS/2
Warp Server for e-business. For OS/2 Warp Server Version 4, we recommend
making the following additions in this order:
• FixPak 42 (or later)
30
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
•
•
•
•
•
Netscape Navigator 4.61 (September, 1999, or later)
Feature Installer 1.2.5 (or later)
Java 1.1.8 (or later)
TCP/IP Version 4.1 (or later)
Updates to TCP/IP 4.1, MPTS, Java, and any other OS/2 Warp Server
functions you plan to use (such as OS/2 LAN server)
These additions can be found at the following Web sites:
• http://www.ibm.com/software/os/warp/swchoice/
Netscape Navigator, Feature Installer, Java, TCP/IP 4.1, and other
enhancements
• http://ps.software.ibm.com
Updates and fixes for all IBM PC software products
• ftp://ftp.hursley.ibm.com
Updates to Java
2.2.2 DHCP server preparation
The first thing you need to do after the DHCP server is fully installed is to set
a password for the administrator. You may also wish to make some changes
to various TCP/IP settings.
Note
TCP/IP Versions 4.1 and 4.21 for OS/2 use Java-based graphical setup
panels. Consequently, you can manage OS/2 Warp Server’s DHCP server
from just about any Web browser. (See Chapter 11, “Managing your
network” on page 411 for details.) However, when you make changes using
the Java graphical panels you should remember the following:
• Scroll bars. Sometimes all the information cannot fit within each window.
You may have to scroll up (or down) to see everything.
• Horizontal adjustments. You may be able to read some information
better if you make horizontal adjustments to some of the header
windows (for example, Option Name).
We suggest that you create backup copies of the following files before
starting:
C:\CONFIG.SYS
C:\MPTN\BIN\SETUP.CMD
C:\MPTN\BIN\MPTSTART.CMD
C:\TCPIP\BIN\TCPSTART.CMD
Chapter 2. Up and Running with DHCP
31
where the C: drive may be different, depending on how you installed OS/2
Warp Server.
Figure 8. [Warp Server for e-business] TCP/IP Version 4.21 configuration folders
To change the TCP/IP administrator password:
1. Open your TCP/IP Shadows folder and then the TCP/IP Configuration
folder (Figure 8).
2. Double-click on Create TCP/IP Administrator Password. Enter your
administrator password twice, and click on OK (Figure 9). Be sure to
remember this password.
Figure 9. [Warp Server for e-business] TCP/IP administrator password window
32
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Then, you can perform a basic setup using the TCP/IP configuration
notebook. (Some of these settings may already be filled in, depending on
whether you entered any TCP/IP settings during installation of OS/2 Warp
Server.)
3. In the TCP/IP Configuration folder, double-click on TCP/IP
Configuration (Local). This starts the TCP/IP Configuration Notebook.
(Figure 10 on page 33 shows the values we used.)
4. Assign a static IP address to your DHCP server. Click on the Network tab
and, with LAN interface 0 highlighted, enter the IP address and subnet
mask for the server. (The first network adapter in your server running
TCP/IP is normally assigned LAN interface 0.) Click on the Manually
Using button, if necessary, and make sure that the Enable Interface
checkbox is checked.
The server’s IP address and subnet mask are part of the Basic choices.
None of the advanced options are needed at this time.
Figure 10. [Warp Server for e-business] TCP/IP configuration notebook
Click on the Routing tab and then Add. (See Figure 11 on page 34, for the
routing details we used.)
Chapter 2. Up and Running with DHCP
33
You may already have a default route defined, which you can change if
needed.
Figure 11. (Warp Server for e-business) Routing in the TCP/IP configuration notebook
5. Click on the Host Names tab and enter your computer’s hostname and
domain name. For example, the hostname could be merlot and the domain
name armonk.cooking.net.
6. Now, click on the Hosts tab at the bottom, and then Add a loopback
address (127.0.0.1) using the hostname localhost.
We chose not to Autostart any services, such as FTPD or TELNETD, at
this time.
7. Click on the General tab and choose your correct time zone. We did not
use Security, because we are not presently providing any of the services
on this page. Also, we are not using SOCKS, Printing or NFS.
8. When you are satisfied with all your entries, pressing the OK button at
bottom of the window and press OK again to continue. Your new
configuration will be saved in CONFIG.SYS and \MPTN\BIN\SETUP.CMD. Shut
down and reboot your server now so that these changes take effect.
34
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
You have now completed the basic setup of TCP/IP 4.21 for your DHCP
server.
Figure 12. [Warp Server for e-business] DHCP Server Services
2.2.3 Basic configuration
After completing basic TCP/IP setup, as shown in the previous section, you
are now ready to create a simple DHCP configuration.
1. Open the TCP/IP Shadows folder again, then DHCP Server Services,
and double-click on DHCP Server Configuration (Figure 12). A DHCP
Server Configuration window should appear to request your TCP/IP
administrator password (Figure 13).
Figure 13. [Warp Server for e-business] Password for DHCP Server Configuration
2. Enter your password and press Enter (or click on OK). Be patient, after
you press Enter, it will take some seconds before the DHCP Server
Configuration window appears (Figure 14 on page 36). You may have to
click on the window to bring it to the foreground. The mouse cursor will
change shape as you move it around the window, indicating the areas that
can be enlarged or reduced in size.
We recommend that you use this DHCP server configuration window to
make any configuration changes, rather than editing DHCP configuration
files directly. This Java program automatically checks your entries as they
Chapter 2. Up and Running with DHCP
35
are made. If you edit the configuration file directly, it can become very
difficult to resolve any errors.
A default configuration file should be loaded when you first open the
DHCP Server Configuration window, but you may wish to create a new file.
(The configuration file is \MPTN\ETC\DHCPSD.CFG. You should keep
backups of this critical file.) You can also load a sample file by selecting
File, then Load Sample.
The remaining steps describe what we did to create a DHCP configuration
for our one subnet network.
Figure 14. [Warp Server for e-business] DHCP Server sample configuration
3. Click on File and then New (see Figure 15 on page 37).
36
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 15. [Warp Server for e-business] DHCP Server new configuration
4. Click on Configure, then Add Subnet. After a few seconds, a notebook
should open. Fill in the values as desired.
5. Click on the DHCP Options tab. As shown in Figure 16 on page 38, we
chose only option 1, the subnet mask (255.255.255.0), and option 15, the
domain name (armonk.cooking.net).
Chapter 2. Up and Running with DHCP
37
Figure 16. [Warp Server for e-business] DHCP Subnet notebook
6. Save your changes [File => Save].
7. Start the DHCP server to check your new configuration. Double-click on
the DHCP Server icon in the DHCP Server Services folder. As shown in
Figure 17 on page 39, you should see a message when the server starts
(in a separate window) that says Server Initialized.
38
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 17. [Warp Server for e-business] DHCP Server initialized
If there are problems, the DHCPSD.LOG file is a good place to look. This text
file usually resides in the root directory. You may change the location of this
file if you wish by double-clicking on DHCP Server in the DHCP Server
Configuration window. A settings notebook appears, as shown in Figure 18 on
page 40. You can enter the full path to a new log file in the Log file name field.
You can also choose which kinds of errors should be logged to this file by
checking or unchecking each checkbox.
Chapter 2. Up and Running with DHCP
39
Figure 18. [Warp Server for e-business] DHCP log file notebook settings notebook
2.2.4 Testing and verification
Now that your DHCP server is running, your DHCP clients should be able to
get IP addresses and parameters from the server without difficulty. We tested
many different DHCP clients with the OS/2 Warp Server DHCP server,
including:
•
•
•
•
•
•
•
•
•
•
•
40
OS/2 Warp 4
Windows 95 and Windows 98
Windows NT Workstation 4.0
Windows 2000
Apple Macintosh
IBM Network Station
AIX
Linux
DOS, Windows 3.1, and Windows for Workgroups
Sun Microsystems Solaris
IBM WorkSpace On-Demand
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• Hewlett-Packard LaserJet 4000 Printer with JetDirect
None of these clients had any problems communicating with the IBM DHCP
server. (You can learn how to configure each of these clients in 2.5, “DHCP
clients” on page 64.)
To determine which IP addresses are being used by your clients, open an
OS/2 command line window and enter the following command on the server:
DADMIN -S | MORE
(You may also wish to check address assignments at each client. Or, you may
wish to examine the DHCP log file on the server.)
You can also use the DADMIN command to perform any of the following tasks:
• Re-initialize the DHCP server. (The server re-reads its configuration file
and adopts any changes.)
• Delete a lease.
• Control server tracing.
• Display client information.
• Display IP address information.
• Display server statistics.
DADMIN Command Syntax
dadmin [-?]
[-v]
[-v]
[-v]
[-v]
[-v]
[-v]
[-v]
[-v]
[-v]
[-v]
[-v]
[-v]
[-v]
[[-h]<host> ] [-f] -d <ipaddress>
[[-h]<host> ] -i
[[-h]<host> ] -bi
[[-h]<host> ] -s
[[-h]<host> ] -t <on/off>
[[-h]<host> ] -bt <on/off>
[[-h]<host> ] -n <intervals>
[[-h]<host> ] -q <ipaddress>
[[-h]<host> ] -p <ipaddress>
[[-h]<host> ] -c <client id>
[[-h]<host> ] -u <User Password>
[[-h]<host> ] -x
[[-h]<host> ] -a
Where:
-?
Displays usage information.
Chapter 2. Up and Running with DHCP
41
42
-v
Enables verbose mode. This mode provides additional information for
debugging purposes.
-f
Forces deletion of the lease without prompting.
-h
Specifies either the IP address or the hostname of the DHCP server. If
no server is specified, the local server is assumed.
-d
Deletes the lease for the specified IP address.
-i
Re-initializes the specified server.
-bi
Re-initializes the specified BINL server.
-s
Displays all IP addresses currently leased.
-t
Specifies server tracing. Use a value of ON to start tracing or OFF to
stop tracing.
-bt
Specifies BINL server tracing. Use a value of ON to start tracing or OFF
to stop tracing.
-n
Requests statistics for the specified server. The value is a decimal
integer indicating the number of intervals from 0 to 100. For example,
a value of three returns a summary record that includes totals
information, the current interval record, and the three most recent
history records. A value of zero returns a summary record of activity
since the last summary.
-q
Requests the IP address information.
-p
Requests the address pool information.
-c
Requests information for one or more clients that match the client ID.
-u
Specifies the administrator password. If the DHCP server is local, this
parameter is not required. If the password contains blanks, you must
enclose the password in double quotes (").
-x
Specifies that the DHCP server is older than the version supplied with
IBM TCP/IP Version 4.21. Access is controlled by the RHOSTS file on
the DHCP server, which requires either a HOSTS file or DNS name
resolution.
-a
Specifies that no administrator password checking is to be performed.
This parameter is used for DHCP servers that are not running on IBM
TCP/IP Version 4.21 or on IBM eNetwork Network Station Manager
TCP/IP for Windows NT.
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
2.3 Windows NT as a DHCP server
This section describes how to set up a simple DHCP server using Microsoft
Windows NT Server 4.0 with Service Pack 5.
Note
Service Pack 5 for Windows NT Server 4.0 resolved earlier DHCP
problems that included:
• Ignored Reservations.
• The inability to display years from 2000 to 2009 in the DHCP
administrators’ program.
• Using DHCPCMD to modify a reservation, but instead it converted it to a
lease.
• The DHCP reconcile could not recover reservations that were out of
scopes.
• The DHCP MIB returned an incorrect OID for the SNMP GetNext
command.
• The DHCP server service could cause an access violation while
administrating locally or remotely.
• Could not obtain a lease with DHCP scopes using multicast addresses.
• The DHCP scope was marked Bad after the installation of Windows NT
Service Pack 3.
• DHCP was disabled when connected to an Internet service provider.
For more detailed information on Service Pack 5, please see:
http://www.microsoft.com/NTServer/Support/faqs/sp5faq.asp
2.3.1 Installation
When you install Windows NT Server 4.0, the DHCP server is not installed by
default. If you have not installed NT’s DHCP server yet, follow these steps:
1. Click on the Start menu. Select Settings and then select Control Panel.
2. Double-click on the Network icon.
3. Click on the Services tab in the Network notebook, then click on the
Add... button.
4. Select Microsoft DHCP Server from the list and then click on OK.
Chapter 2. Up and Running with DHCP
43
5. When prompted, enter the path to the Windows NT files (Windows NT
server CD-ROM) and click on Continue to copy the files and finish the
installation.
6. If you had a Service Pack installed on the server before you installed the
DHCP server then you must re-install the Service Pack.
2.3.2 DHCP server preparation
We assume a network with one subnet for this example. All clients attached to
the network were previously configured using static IP addresses, including
static DNS server addresses. The following changes will be made in our
sample network:
• Adding a DHCP server to the network
• Configuring all clients to obtain IP addresses, subnet mask, router
address, and name server information from the DHCP server
The following values represent our sample network configuration:
Network ID
Subnet Mask
Domain Name
DNS Server
DHCP Server
192.168.8.0
255.255.255.0
bellevue.cooking.net
192.168.8.10
192.168.8.10
You should plan your own network configuration before setting up any DHCP
server. You should take the following requirements into consideration:
• Non-DHCP clients must have static IP addresses, and these must be
excluded from the IP address pool.
• The DHCP server must be assigned a static IP address. It cannot be a
DHCP client to itself.
• You should determine the range of IP addresses (the scope) that can be
assigned to DHCP clients.
• Any other DHCP option values (such as name servers) should also be
defined for the clients.
• The duration of the IP address lease must be chosen.
Much like 2.2, “OS/2 Warp Server for e-business as a DHCP server” on page
30, we chose to have our DHCP server handle the private Class C addresses
192.168.8.1 through 192.168.8.254, with 192.168.8.1 through 192.168.8.10
excluded from this pool in order to reserve a small set of static addresses.
One of those reserved addresses, 192.168.8.10, was assigned to the DHCP
44
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
server itself (which happened to also act as a DNS server). The subnet mask
was 255.255.255.0, and the lease duration was set to one day.
We also chose the following option values for our sample network:
Option 003: Router
Option 006: DNS Server
Option 015: Domain Name
192.168.8.1
192.168.8.10
bellevue.cooking.net
Before you start configuring the DHCP server, you must make sure that
TCP/IP is set up and configured correctly. In particular, the DHCP server
must have a static IP address. To configure TCP/IP manually, follow these
steps:
1. Open the Control Panel and then open the Network icon.
2. Click on the Protocols tab.
3. Select TCP/IP Protocol and click on Properties.
4. Click on the IP Address tab. Check Specify an IP address and type in
the server’s IP address, subnet mask, and default gateway address.
5. Click on the DNS tab and type in the hostname and domain name for the
server.
6. Under DNS Service Search Order, click on Add.
7. In the DNS server box, type in the IP address for your server (assuming it
is also acting as a DNS server) and click on Add.
8. Click on OK to close the notebook.
9. Click on OK to close the Network notebook.
2.3.3 Basic configuration
After assigning a static IP address to the DHCP server, as shown in the
previous section, you are now ready to create a simple DHCP configuration.
1. By default, the Microsoft DHCP server service starts automatically when
the computer is started. If the DHCP server is not running, open the
Services icon in the Control Panel, select the Microsoft DHCP Server
from the service list, and click on the Start button. You can also start the
DHCP server from the command line, using the following command:
NET START DHCPSERVER
2. Start the DHCP manager. Click on the Start menu, select Programs, then
select Administration Tools (Common), and click on DHCP Manager.
Chapter 2. Up and Running with DHCP
45
3. You can manage both local and remote DHCP servers using this utility. If
you want to access a remote DHCP server, select Server and then select
Add from the menu. The Add DHCP Server to Server List window will
appear. Enter the address(es) of the remote DHCP server(s) you wish to
manage and click on OK.
4. Next, create the scope for the subnet. In the DHCP manager window,
highlight the Local Machine (or the remote server you wish to manage),
click on the Scope menu, and then select Create. Fill in the following
information (see Figure 19 on page 47):
Start Address
Starting IP address for the scope (such as
192.168.8.1).
End Address
Ending IP address for the scope (such as
192.168.8.254).
Subnet Mask
Subnet mask assigned to the clients (such as
255.255.255.0).
Exclusion Range You can exclude any static IP addresses from the
scope using this option. Type in the start address and
the end address, and then click on the Add button.
These addresses will not be assigned to DHCP
clients.
46
Lease Duration
Duration of the address lease. Click on the Limited To
button and then type in the number of days, hours, and
minutes to set the lease duration.
Name
A name for the DHCP scope. This name appears in
the DHCP manager solely for your convenience.
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 19. [Windows NT] Creating a scope
5. After filling in all the information, click on OK to save the configuration. A
message will appear to remind you that the scope has not been activated
yet. If you click on Yes, you will activate the scope immediately. However,
you should not activate a new scope until you have specified its DHCP
option values, so click on No.
6. Now, you should set the option values passed to DHCP clients. In the
DHCP Servers list, highlight the scope you just created, click on the
DHCP Options menu, then select Scope.
7. In the Unused Options list, choose an option to be added and click on the
Add button. The option you selected is moved to the Active Options list.
(We configured options 3, 6, and 15, as shown in Figure 20.)
Figure 20. [Windows NT] Adding DHCP Options
Chapter 2. Up and Running with DHCP
47
8. To define a value for each option, highlight an option in the Active Options
list and click on Value. Enter a value in the appropriate field for the option,
as shown in Figure 21. Repeat for each Active Option.
Figure 21. [Windows NT] Entering value for option 15
Some options can take on multiple values, such as option 6 (DNS
Servers). For these options, click on Edit Array, enter one of the values,
then click on Add. You can enter more values, one by one, and click on
Add to build the full list of values for the option (see Figure 22). When you
finish building the list, click OK.
Figure 22. [Windows NT] Creating address list for option 3
48
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Note
You cannot choose option 1 (subnet mask) from the DHCP options list in
Windows NT server. The value is taken from the scope property
notebook (see Figure 19 on page 47).
When you have finished defining all the option values, click on OK.
9. If the scope is not active (if a small icon representing the scope in the list
is grayed out), click on Scope and then click on Activate. The icon should
turn yellow.
2.3.4 Testing and verification
1. Start your DHCP clients and check whether they receive address
assignments. (We tested Windows 95, Windows NT Workstation 4.0, and
OS/2 Warp 4 as DHCP clients. All DHCP clients received TCP/IP
configuration information through the Windows NT DHCP server properly.)
2. To check for active leases from the server, highlight the scope you want to
view, select Scope, then select Active Lease. The Active Leases window
(Figure 23) will appear and list active leases.
Figure 23. [Windows NT] DHCP Active Leases
Chapter 2. Up and Running with DHCP
49
Select any system in the client list, then click Properties to view more
information about an individual DHCP client. (See Figure 24 on page 50.)
Figure 24. [Windows NT] DHCP Client Properties
2.4 AIX as a DHCP server
Note
Before configuring AIX Version 4.3.3 as a DHCP server, ensure that the
file set bos.net.tcp.server is installed. To verify this, use the command:
lslpp -l “bos.net.tcp.*”
To install this file set, you will need AIX CD-ROM 1, and you may be
prompted for further CD-ROMs.
The DHCP configuration for AIX is held in the file /etc/dhcpsd.cnf. This file
can be created and maintained either manually by using a text editor or by
using the dhcpsconf command.
2.4.1 Manually configuring the DHCP server
The configuration file /etc/dhcpsd.cnf is a flat text file that can be maintained
with any text editor, for example vi. For smaller locations, it is faster and
easier to maintain the configuration file manually than using the GUI
application.
A sample file including documentation is copied when AIX is installed.
50
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
2.4.2 Configuring DHCP server through dhcpsconf
dhcpsconf is an X-windows GUI that lets you create and maintain
configuration files and also allows you to start, stop and retrieve statistics
from a running server. dhcpsconf also allows you to create a master definition
file that can contain the configurations for all your AIX DHCP servers, and
then generate individual dhcpsd.cnf files as required. Figure 25 on page 52
shows the sample environment we are going to configure. It consists of two
token-ring networks. The two networks are interconnected with one Windows
NT Server Version 4 machine, mustard, that acts as a router. The AIX system,
buffet, acts a DHCP server.
There is a DNS available to all networks, and its IP address is 192.168.6.2.
Devices on the 192.168.6.0 subnet include LAN workstations that must
resolve NETBIOS names to IP addresses; so, an NBNS is available as
192.168.6.5. (See Chapter 5, “Integrating file and print services” on page 227
for information on NETBIOS name resolution and NBNS.)
In this network, the server buffet has DHCP configured to serve the subnet
192.168.6.0. Another DHCP server could be setup on the same network to
serve the subnet 192.168.7.0 and would see DHCP requests from the
192.168.6.0 subnet, but would ignore them because it contains no
configuration for that subnet.
Chapter 2. Up and Running with DHCP
51
8 8.0
192.16
192.168 8.10
mustard
NBNS
DNS
192.168 6.2
192.168 6.5
192.168 6.2
192.1
68 6.0
192.168 6.4
buffet
192.168 9.4
192.168 9.0
Figure 25. Sample environment
To start the DHCP server configuration, execute dhcpsconf from a terminal
window. The GUI will be open, with an empty configuration as shown in
Figure 26 on page 53. We can now begin to create our master definition file.
52
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 26. [AIX] DHCP Server Configuration
Add the server by highlighting Server in the Key List window and then
pressing the Add button underneath the Key List window.
A panel like the one in Figure 27 on page 54 will be displayed. Fill in the
Name and Machine name with the hostname of the server you want to
configure (for clarity, we have entered the full hostname). The field names are
self-explanatory.
Chapter 2. Up and Running with DHCP
53
Figure 27. [AIX] Adding a DHCP Server
Change view so that we are only working with one server by selecting [Server
– View by ...]. A selection window displaying servers in the master file will be
shown. Select the server you want to work with. Notice the <Master File>
option in Figure 28. Choosing this option allows you to view the configuration
for all the servers contained within the master file.
Figure 28. [AIX] Choosing a DHCP Server
We’ve chosen to modify the configuration for the server buffet. The first thing
we will do is add some global options for our network. From the Option List
54
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
window, we add a definitions for the Domain Name (armonk.cooking.net) and
a DNS (192.168.6.2), as shown in Figure 29 and Figure 30.
Figure 29. [AIX] Adding a domain name
Figure 30. [AIX] Adding a DNS
Next, we will add the 192.168.6 network. From the Key List window, add a
Network. The panel shown in Figure 31 on page 56 will be displayed.
As you add any item, the Name field will be filled in with a simple description
and a number. The number is incremented each time you add an item. This is
used to ensure that all the names are unique. You can change the name to
whatever you wish, as long as it is unique within your master file.
Modify the Network Address field to contain the IP address for the network.
This must be a full address, for example 192.168.6.0, not 192.168.6.
In the Number of bits, or low end of a range field, enter the number of bits that
define the subnet mask for the network. If your mask is 255.255.255.0, then
you have a 24-bit mask. Alternatively, you can enter the standard mask.
Chapter 2. Up and Running with DHCP
55
Figure 31. [AIX] Adding a network
Note that as you add options to your DHCP server, they are also being added
to the Options List and Key List windows. This allows them to be added as
required to the DHCP server file window by simply highlighting them and
pressing Add.
Now you can add the first of our subnets. From the Key List window, add a
Subnet. The panel shown in Figure 32 on page 57 will be displayed.
In the Subnet Address field, enter the full IP address for the subnet, for
example 192.168.6.0. In addition, add the low and high end of the range of
addresses that will be served for this subnet. If you leave these fields blank,
all addresses within the subnet will be served.
56
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 32. [AIX] Adding a subnet
Options can now be added to the subnet just created. Highlight the subnet in
the DHCP server file window, and choose to add options from the Option List
window. As shown in Figure 33, add a default router for the 192.168.6.0
subnet.
Figure 33. [AIX] Adding a default router
With the addition of other subnets and options, we end up with the
configuration for the buffet server that looks like Figure 34 on page 58.
Chapter 2. Up and Running with DHCP
57
Figure 34. [AIX] Completed Server Configuration
If you have any other DHCP servers you wish to configure, then select them
and configure them now. Once all the servers have been configured, select
[Server =>View By ... => <Master File>]. Expand all the items in the DHCP
Server file window to see your complete configuration, as shown in Figure 35
on page 59.
58
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 35. [AIX] Final configuration
Select [File => Save As] and change the Selection field to /etc/masterfile.cnf
before clicking on OK.
Chapter 2. Up and Running with DHCP
59
Figure 36. [AIX] Saving the master file with [File – Save As]
To build the dhcpsd.cnf file for an individual server, select [Server =>
Save File for]. The server selection panel will be displayed.
Assuming we are working at the machine buffet and we want to build its
configuration file, we select buffet.armonk.cooking.net and click on OK.
The DHCP Save File Name panel is displayed. Change the Selection field to
/etc/dhcpsd.cnf, then click on OK.
Figure 37. [AIX] Saving an individual server file with [Server – Save File for]
60
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Select [File => Quit] to exit dhcpsconf.
After configuration, /etc/masterfile.cnf is shown in Figure 38.
###
###
###
###
###
###
###
###
"buffet.armonk.cooking.net"
"buffet.armonk.cooking.net"
"buffet.armonk.cooking.net"
"buffet.armonk.cooking.net"
"buffet.armonk.cooking.net"
"buffet.armonk.cooking.net"
"buffet.armonk.cooking.net"
"buffet.armonk.cooking.net"
numLogFiles 0
logFileSize 0
logFileName
leaseTimeDefault 0 minute
leaseExpireInterval 0 minute
supportBOOTP No
supportUnlistedClients No
## DEFAULTIPADDRESS 0.0.0.0
## DEFAULTHOSTNAME
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
### "buffet.armonk.cooking.net"
"option 15 armonk.cooking.net"
"option 5 192.168.6.2"
"subnet 192.168.6.0 192.168.6.90-192.168.6.100"
"{"
"option 3 192.168.6.1"
"option 1 255.255.255.0"
"}"
"network 192.168.6.0 24"
"numLogFiles 10"
"logFileSize 100"
"logFileName /usr/log/dhcpsd.log"
"logItem SYSERR"
"logItem OBJERR"
"logItem PROTERR"
"logItem WARNING"
"logItem EVENT"
"logItem ACTION"
"logItem INFO"
"logItem ACNTING"
"logItem TRACE"
"leaseTimeDefault 24 hour"
"leaseExpireInterval 5 minute"
"supportBOOTP No"
"supportUnlistedClients Yes"
"IPADDRESS 0.0.0.0"
"HOSTNAME buffet"
Figure 38. [AIX] The masterfile.cnf file
The dhcpsd.cnf file created for buffet with the [Server => Save File for] menu
options is shown in Figure 39 on page 62.
Chapter 2. Up and Running with DHCP
61
option 15 armonk.cooking.net
option 5 192.168.6.2
subnet 192.168.6.0 192.168.6.90-192.168.6.100
{
option 3 192.168.6.1
option 1 255.255.255.0
}
network 192.168.6.0 24
numLogFiles 10
logFileSize 100
logFileName /usr/log/dhcpsd.log
logItem SYSERR
logItem OBJERR
logItem PROTERR
logItem WARNING
logItem EVENT
logItem ACTION
logItem INFO
logItem ACNTING
logItem TRACE
leaseTimeDefault 24 hour
leaseExpireInterval 5 minute
supportBOOTP No
supportUnlistedClients Yes
## DEFAULTIPADDRESS 0.0.0.0
## DEFAULTHOSTNAME buffet
Figure 39. [AIX] The dhcpsd.cnf file
2.4.3 Disabling BOOTP support in inetd
Before the DHCP server can be started, support for the BOOTP server must
be removed from inetd because both BOOTP and DHCP use the same ports
(67 and 68) to receive and send data. With BOOTP running through inetd,
port 67 is unavailable, and dhcpsd will not be able to start.
Issue the command:
grep bootps /etc/inetd.conf
If the line is commented (that is, the first character is a #), then BOOTP is
disabled, and the DHCP server can be started.
If it isn’t, you can manually edit the /etc/inetd.conf file, and comment out the
line.
Alternatively, start SMIT with a smit inetd command. Then select
[inetd Subservers => Remove an inetd Subserver], and finally choose
bootps udp from the list of Subservers that may be removed.
inetd must be restarted after modification of the configuration file. Assuming
that inetd is running under SRC control, issue the command:
62
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
refresh -s inetd
If it isn’t, then issue the command:
kill -1 inetd_pid
Where inetd_pid is the pid of inetd.
2.4.3.1 Starting the DHCP server
The DHCP server subsystem can be started in several ways.
• From within dhcpsconf, select [Operations => Start]. Select the name of
the DHCP server you want started.
Select OK.
• From a command prompt, issue the command:
smit dhcpsd
Select Start using the dhcp Subsystem.
Select NOW.
• From a command prompt, issue the command:
startsrc -s dhcpsd
2.4.3.2 Stopping the DHCP server
The DHCP server subsystem can also be stopped in several ways. All of
these methods only stop the DHCPCD subsystem; they do not affect the rest
of AIX:
• From in dhcpsconf, select [Operations => Stop].
Select the name of the DHCP server you want stopped.
Select OK.
• From a command prompt, issue the command:
smit dhcpsd
Select Stop Using the dhcpsd Subsystem.
Select NOW.
• From a command prompt, issue the command:
stopsrc -d dhcpsd
Chapter 2. Up and Running with DHCP
63
2.5 DHCP clients
Because a DHCP server, such as OS/2 Warp Server for e-business, supports
open Internet standards, it can service a wide variety of clients. Most large
TCP/IP networks include many different kinds of systems running a wide
variety of operating systems. You may be confronted with any or all of the
types of systems described in this section.
This section explains how to configure several different clients to receive
TCP/IP address assignments from a DHCP server.
2.5.1 OS/2 Warp 4
OS/2 Warp 4 includes TCP/IP Version 4.0 as a standard feature. You may
wish to upgrade OS/2 Warp 4’s TCP/IP capabilities to Version 4.1, available
from the IBM Software Choice Web site, or to TCP/IP Version 4.21 for OS/2,
which is a component of OS/2 Warp Server for e-business.
Note
You can download IBM Software Choice packages, FixPaks, and other
enhancements from the Internet. Please refer to Appendix B, “Where Is It?
Internet and IBM Intranet Web Sites” on page 475 for more information.
Both TCP/IP versions support DHCP, but they differ in the way DHCP is
activated. (Skip ahead to 2.5.1.3, “OS/2 Warp 4 DHCP client configuration”
on page 67, if you plan to use TCP/IP 4.0.)
2.5.1.1 Installing TCP/IP 4.21
At a minimum, you must have the following software available prior to
installing TCP/IP 4.2.1:
• OS/2 Warp 4 with TCP/IP 4.0
• MPTS Version 5.5 (or later)
• Netscape Navigator 4.61 (September, 1999, or later)
• Feature Installer 1.2.5 (or later)
Feature Installer provides plug-ins for Netscape Navigator that are used to
install Java Version 1.1.8 and TCP/IP Version 4.21.
• Java Version 1.1.8 (or later)
Only the Java runtime is required.
To install TCP/IP Version 4.21 on OS/2 Warp 4, follow these steps:
64
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
1. Install OS/2 Warp 4 with its built-in TCP/IP 4.0 client services (and other
services as needed).
2. Install Netscape Navigator 4.61 (September, 1999, or later) for OS/2
Warp.
3. Install FixPak 12 (or later) for OS/2 Warp 4. Make sure you shutdown and
reboot after installing the FixPak.
4. Install any other FixPaks (for example, File/Print client) as needed.
5. Install Feature Installer 1.2.5 (or later). Shutdown and reboot.
6. Install Java 1.1.8 (or later; we recommend installing Java 1.1.8). Only the
runtime is required, but you may wish to install other Java functions.
7. Install MPTS Version 5.5 (or later). This is available on the OS/2 Warp
Server for e-business CD.
• When asked whether to install a more recent version of MPTS, select
Yes. Your current MPTS configuration will be preserved, but MPTS
drivers and protocols will be updated.
• After installation of the most recent version of MPTS, click on Exit.
• When asked whether to update your CONFIG.SYS file, click on OK.
• Shut down and restart your computer after exiting MPTS.
8. Install TCP/IP Version 4.21 for OS/2.
• Run install.cmd in the TCPAPPS directory. This starts Netscape
Navigator.
• Select the Advanced Path option for this installation.
• Ensure that there is a tick in the check box next to TCP/IP Base
Applications and that the target drive is C.
• Allow the program to start copying the files.
• When the copying has finished, shut down and restart your computer
after exiting the installation program.
9. Apply the latest TCP/IP 4.21, MPTS 5.5, and Java fixes.
Note: You do not need to provide Java fixes if Java 1.1.8 was installed.
The installation of TCP/IP Version 4.21 for OS/2 creates a TCP/IP Shadows
folder on your desktop.
2.5.1.2 Installing TCP/IP 4.1
At a minimum, you must have the following software available prior to
installing TCP/IP 4.1:
Chapter 2. Up and Running with DHCP
65
• OS/2 Warp 4 with TCP/IP 4.0
• MPTS Version 5.3 (or later)
• Netscape Navigator 4.61 (September, 1999, or later)
• Feature Installer 1.2.5 (or later)
Feature Installer provides plug-ins for Netscape Navigator that are used to
install Java Version 1.1.8 and TCP/IP Version 4.1.
• Java Version 1.1.8 (or later)
Only the Java runtime is required.
To install TCP/IP Version 4.1 on OS/2 Warp 4, follow these steps:
1. Install OS/2 Warp 4 with its built-in TCP/IP 4.0 client services (and other
services as needed).
2. Install Netscape Navigator 4.61 (September, 1999, or later) for OS/2
Warp.
3. Install FixPak 12 (or later) for OS/2 Warp 4. Make sure you shutdown and
reboot after installing the FixPak.
4. Install any other FixPaks (for example, File/Print client) as needed.
5. Install Feature Installer 1.2.5 (or later). Shutdown and reboot.
6. Install Java 1.1.8 (or later). Only the runtime is required, but you may wish
to install other Java functions.
7. Use Netscape Navigator’s Java switching utility to switch to Java 1.1.
Open the Netscape folder and double-click on Java Version Selection for
Netscape Navigator. Select the appropriate Java version and click OK.
8. Install TCP/IP Version 4.1 for OS/2. A new version of MPTS (Version 5.3)
and new TCP/IP applications will be installed.
• When asked whether to install a more recent version of MPTS, select
Yes. Your current MPTS configuration will be preserved, but MPTS
drivers and protocols will be updated.
• After installation of the most recent version of MPTS, click on Exit.
• When asked whether to update your CONFIG.SYS file, click on OK.
• Do not shutdown or restart your computer after exiting MPTS. Continue
with the rest of TCP/IP Version 4.1 installation.
• After installation of MPTS, Netscape Navigator starts. Follow the
instructions provided to complete installation. Shutdown and reboot
your PC when TCP/IP Version 4.1 installation has finished.
66
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
9. Apply the latest TCP/IP 4.1, MPTS 5.3, and Java fixes.
Note: You do not need to provide Java fixes if Java 1.1.8 was installed.
The installation of TCP/IP Version 4.1 for OS/2 creates a TCP/IP Shadows
folder on your desktop.
2.5.1.3 OS/2 Warp 4 DHCP client configuration
DHCP client configuration for OS/2 Warp 4 depends on the version of TCP/IP
you are running.
• TCP/IP Versions 4.21 and 4.1 for OS/2
1. Open the TCP/IP Shadows folder, then the TCP/IP Configurations
folder. Double-click on the TCP/IP Configuration (Local) icon, or just
type TCPCFG2 at an OS/2 command prompt. The TCP/IP Configuration
notebook appears.
2. Click on the Network tab. Highlight the appropriate LAN interface from
the list (normally LAN interface 0) and make sure the Enable
Interface checkbox is checked. Then, select Automatically, Using
DHCP, as shown in Figure 40 on page 68. Click on OK to close the
TCP/IP configuration notebook and save this change.
Chapter 2. Up and Running with DHCP
67
Figure 40. [OS/2 Warp] Enabling a DHCP client with TCP/IP 4.1
3. Restart your computer to activate the DHCP client function. You can
view the current IP configuration by using the DHCP client monitor
utility. Open the OS/2 System folder, then the System Setup folder.
Double-click on the DHCP Monitor icon. (Or, type DHCPMON at an OS/2
command prompt.) Select [View => Details] for more information on
the lease assignment. (See Figure 41 on page 69.)
Note
With TCP/IP Version 4.21 for OS/2 you may release an IP address by
typing DHCPMON /RELEASE at an OS/2 command prompt. You may also type
DHCPMON /RENEW at a command prompt to renew a lease.
68
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 41. [OS/2 Warp] DHCP Client Monitor using TCP/IP 4.1
You can also view all the parameters received from the DHCP server by
selecting Current Configuration from the Configure menu.
DHCP Client Monitor can obtain or release an IP address assignment
from any DHCP server. Select [Actions => Release lease] to release
an IP address and select [Actions => Request lease] to obtain an IP
address. (You may wish to test a new DHCP server using the DHCP
Client Monitor.)
• TCP/IP 4.0
1. Open the Programs folder then the TCP/IP Internet (LAN) folder.
Double-click on TCP/IP Configuration (LAN). (Or, just type TCPCFG at
an OS/2 command prompt.) The TCP/IP Configuration notebook
appears.
2. Click on the Network tab. Highlight the appropriate LAN interface from
the list (normally LAN interface 0) and make sure the Enable
interface checkbox is checked. Then, select Automatically, Using
DHCP, as shown in Figure 42 on page 70.
Chapter 2. Up and Running with DHCP
69
Figure 42. [OS/2 Warp] Enabling a DHCP client using TCP/IP 4.0
3. Close the TCP/IP configuration notebook to save this change.
4. Shutdown and restart your PC to initiate the DHCP client. You can view
the current IP configuration by using the DHCP Client Monitor. Open
the OS/2 System folder, then the System Setup folder. Double-click on
the DHCP Monitor object. (Or, just type DHCPMON at an OS/2 command
prompt.) Select [View => Details] for more information. (See Figure 43
on page 71.)
70
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 43. [OS/2 Warp] DHCP Client Monitor window
You can also view all the parameters received from the DHCP server by
selecting the Current Configuration button below the Message log
window (see Figure 44).
Figure 44. [OS/2 Warp] Current configuration window
Chapter 2. Up and Running with DHCP
71
2.5.2 Windows 95 and Windows 98
The TCP/IP Properties pages for both Windows 95 and Windows 98 are
identical. Before running your Windows 95 or Windows 98 workstation as a
DHCP client, make sure that you have installed the following components:
• TCP/IP protocol
• An appropriate network adapter device driver
1. To set up your Windows 95 or Windows 98 computer as a DHCP client,
you need to configure the TCP/IP protocol. Open the Network icon. To
open the Network icon, select [Start => Settings => Control Panel].
Double-click on the Network icon.
2. In the Network notebook, highlight TCP/IP Protocol and click on the
Properties button. The TCP/IP properties notebook opens up. Click on
the IP Address tab and check Obtain an IP address automatically, as
shown in Figure 45.
Figure 45. [Windows 98] Enabling DHCP client
3. Click OK to finish the Network settings and reboot your system. Now, the
Windows 95 computer should be able to obtain an IP address from the
DHCP server.
4. You can view the current IP configuration by using the WINIPCFG command
at the command prompt. This will show you a current IP address and other
72
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
TCP/IP configurations, as shown in Figure 46 on page 73. This also
enables you to release and renew the IP address by clicking the buttons.
Figure 46. [Windows 98] IP Configuration window
2.5.3 Windows NT Workstation 4.0
Before running your Windows NT Workstation as a DHCP client, make sure
that you have installed the following components:
• TCP/IP protocol
• An appropriate network adapter device driver
To set up the Windows NT Workstation as a DHCP client, follow the steps
below:
1. Open the Network icon. To open the Network icon, select [Start Settings
=> Control Panel]. Double-click on the Network icon.
2. Click on the Protocols tab, highlight TCP/IP Protocol in the Network
Protocol list, and click on the Properties button. The TCP/IP properties
notebook opens up. To set your computer as a DHCP client, click on the IP
Address tab and check Obtain an IP address from a DHCP server, as
shown in Figure 47 on page 74.
Chapter 2. Up and Running with DHCP
73
Figure 47. [Windows NT] Enabling a DHCP client
3. Click OK to finish the Network settings and reboot your system. Now your
Windows NT Workstation should be able to obtain an IP address from the
DHCP server.
You can view the current IP configuration by using the IPCONFIG utility. Type:
IPCONFIG /ALL
at the command prompt. This will show you the current IP configuration for
the operating system and the network adapter, as shown in Figure 48 on
page 75.
74
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
C:\>ipconfig /all
Windows NT IP Configuration
Host Name. . . . . . . . . :
DNS Servers. . . . . . . . :
Node Type. . . . . . . . . :
NetBIOS Scope ID . . . . . :
IP Routing Enabled . . . . :
WINS Proxy Enabled . . . . :
NetBIOS Resolution Uses DNS:
No
No
No
Token-Ring adapter IBMTOK1:
Description. . . . .
Physical Address . .
DHCP Enabled . . . .
IP Address . . . . .
Subnet Mask. . . . .
Default Gateway. . .
DHCP Server. . . . .
Lease Obtained . . .
Lease Expires. . . .
IBM Turbo 16/4 Token-Ring ISA Adapter
00-04-AC-FF-CB-3C
Yes
192.168.8.12
255.255.255.0
192.168.8.1
192.168.8.10
Monday, October 04, 1999 10:36:31 AM
Tuesday, October 05, 1999 10:36:31 AM
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
:
:
:
:
:
:
:
:
:
burger.bellevue.cooking.net
192.168.8.10
Broadcast
C:\>
Figure 48. [Windows NT] Displaying current configuration
To give up the current IP lease, type:
IPCONFIG /RELEASE
at the command prompt.
To obtain a new IP address, or to update options and lease time, type:
IPCONFIG /RENEW
at the command prompt.
2.5.4 Windows 2000
The version of Windows 2000 that we tested with was Release Candidate 2
(5.00.2128).
Before running your Windows 2000 Workstation computer as a DHCP client,
make sure that you have installed the following components:
• TCP/IP protocol
• An appropriate network adapter device driver
Chapter 2. Up and Running with DHCP
75
To set up the Windows 2000 computer as a DHCP client, follow the steps
below:
1. Open the Network and Dial-up Connections folder. To open this folder,
select [Start => Settings => Network and Dial-up Connections].
Double-click on the Local Area Connection icon.
2. Click on the Properties button. Scroll down the list and highlight the
Internet Protocol (TCP/IP). It must have a tick in the checkbox, since this
indicates that this protocol is used by this machine.
3. Click on the Properties button. The Internet Protocol (TCP/IP) Properties
notebook opens up. To set your computer as a DHCP client, select Obtain
an IP address automatically, as shown in Figure 49. You may also select
Obtain DNS Server address automatically. If you wish to configure
some of the Advanced features for TCP/IP, then click on the Advanced
button.
Figure 49. [Windows 2000] Enabling a DHCP client
4. Click OK to finish the network settings and reboot your system. Now your
Windows 2000 should be able to obtain an IP address from the DHCP
server.
You can view the current IP configuration by using the IPCONFIG utility. Type:
IPCONFIG /ALL
76
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
at the command prompt. This will show you the current IP configuration for
the operating system and the network adapter, as shown in Figure 50 on
page 77.
C:\>ipconfig /all
Windows 2000 IP Configuration
Host Name. . . . . . . . . :
DNS Servers. . . . . . . . :
Node Type. . . . . . . . . :
NetBIOS Scope ID . . . . . :
IP Routing Enabled . . . . :
WINS Proxy Enabled . . . . :
NetBIOS Resolution Uses DNS:
No
No
No
Token-Ring adapter IBMTOK1:
Description. . . . .
Physical Address . .
DHCP Enabled . . . .
IP Address . . . . .
Subnet Mask. . . . .
Default Gateway. . .
DHCP Server. . . . .
Lease Obtained . . .
Lease Expires. . . .
IBM Turbo 16/4 Token-Ring ISA Adapter
00-04-AC-FF-CB-3C
Yes
192.168.8.12
255.255.255.0
192.168.8.1
192.168.8.10
Monday, October 04, 1999 10:36:31 AM
Tuesday, October 05, 1999 10:36:31 AM
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
:
:
:
:
:
:
:
:
:
burger.bellevue.cooking.net
192.168.8.10
Broadcast
C:\>
Figure 50. [Windows 2000] Displaying current configuration
To give up the current IP lease, type:
IPCONFIG /RELEASE
at the command prompt.
To obtain a new IP address, or to update options and lease time, type:
IPCONFIG /RENEW
at the command prompt.
2.5.5 Apple Macintosh
The Apple Macintosh was one of the first microcomputers to feature built-in
networking, and its designers have focused on simplicity to help make the
system easy to set up and use on a network. The Macintosh can easily
receive its IP address assignment from any DHCP server.
We tested an Apple Macintosh running MacOS Version 8 with its built-in
Open Transport TCP/IP software. (Open Transport Version 1.1.1 or later
Chapter 2. Up and Running with DHCP
77
should be installed on previous versions of MacOS to support DHCP.) Either
the Internet Setup Assistant or the TCP/IP Control Panel (Figure 51 on page
78) can be used to select DHCP.
Figure 51. [Macintosh] Selecting the TCP/IP control panel
If you select the TCP/IP control panel a window appears, as shown in Figure
52.
Figure 52. [Macintosh] Choosing the DHCP Server option
78
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Simply select Using DHCP Server and close the window to activate the
Macintosh’s DHCP client support. (Click on Save, as shown in Figure 53 on
page 79, when prompted.) After a system restart, you can revisit this window
to examine the address assignment received from the DHCP server.
Figure 53. [Macintosh] Saving TCP/IP configuration changes
To further integrate your Macintoshes into your network, you may wish to
consider two additional products:
• IBM LAN Server for Macintosh allows Warp Server to act as an
AppleShare file and print server. The Macintosh’s built-in Chooser can be
used to access shared drives and printers on Warp Server without adding
any additional software to the clients. In addition, Warp Server’s Postscript
translation services (Advanced Printing) can accept Macintosh Postscript
print jobs and format them for PC printers in PCL, IBM PPDS, or AFP
formats, providing Macintosh users access to the same printers as PC
users. For more information on LAN server for Macintosh, please consult
the redbooks available at http://www.redbooks.ibm.com.
• DAVE, produced by Thursby Software Systems, is client software for
Macintoshes designed to communicate with Warp Server, Windows NT,
Windows 95, Windows 98, IBM DOS LAN Services, and other
NetBIOS-based networks. However, DAVE must be installed on each
client. For more information on DAVE, please visit http://www.thursby.com
on the Internet.
2.5.6 IBM Network Station
We used an IBM Network Station computer (NC) model 1000, to test the
BOOTP client function and the DHCP client function of the DHCP server with
TCP/IP Version 4.21.
2.5.6.1 BOOTP
First, we tested the BOOTP function. The NC was configured so that DHCP
was disabled, with BOOTP as the first choice for booting. The configuration
was saved, and the NC was powered off.
Chapter 2. Up and Running with DHCP
79
For this test, we needed the TFTP daemon and the portmapper running on
the server. The NC had to download its kernel file and it used TFTP for that
task. We configured these using TCPCFG2 (or by double-clicking on the TCP/IP
Configuration object). As shown in Figure 54, we selected the Autostart
page, chose inetd, tftpd, and portmapper. If you have any problems selecting
the inetd service, this is not a problem, provided the other two services
autostart.
Figure 54. [Warp Server for e-business] TCP/IP Configuration Notebook: Autostart service
We also decided to start the syslog daemon, as shown in Figure 55.
80
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 55. [Warp Server for e-business] TCP/IP Configuration Notebook: Syslog daemon
It is useful to log activity on your servers, especially when you are using
TFTP. You may wish to click the Minimize box, rather than leave some
screens to start open as we did.
Configuring the DHCP server on OS/2 is fairly simple. Start the DHCP server
configuration (as discussed in 2.2.3, “Basic configuration” on page 35) and
highlight the subnet we created earlier. Select Configure and then Add
Client. A notebook opens up, as shown in Figure 56 on page 82.
Chapter 2. Up and Running with DHCP
81
Figure 56. [Warp Server for e-business] Client Definition settings notebook
We filled it in with our chosen values on the Client Definition page. Note that
the client ID is the MAC address of the device in question. On the DHCP
Options page, we chose option 67, Boot File Name, so that we could boot
the NC. Note that no path is listed for the boot file, as shown in Figure 57 on
page 83.
82
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 57. [Warp Server for e-business] DHCP Options settings notebook
We configured TFTP security for this client (see Figure 58 on page 84), and
made the home directory C:\TFTPBOOT. The directory tree for the NC is
located in the C:\TFTPBOOT directory. The following line was placed in the
C:\MPTN\ETC\TFTPPAUTH file:
C:\TFTPBOOT RW 192.168.6.17
Chapter 2. Up and Running with DHCP
83
Figure 58. (Warp Server for e-business) TCP/IP Configuration Notebook - TFTPD security
Note that the IP address matches the IP address reserved for the NC when
we performed the Add Client, as shown in Figure 56 on page 82.
Options 1, Subnet Mask, and 15, Domain Name, are inherited from the
subnet settings. Now turn on the BOOTP client support by double-clicking on
the DHCP Server, choosing the Miscellaneous page, and checking Support
BOOTP clients, as shown in Figure 59 on page 85.
84
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 59. [Warp Server for e-business) Support BOOTP settings notebook
After saving the notebook, the DHCP server can be re-initialized by starting
DHCP Server Management and choosing Re-initialize. From an OS/2
command line, we executed the following command:
DADMIN -S | FIND /I "-"
to see the in-use addresses, then we turned on the NC. When going back and
adding options like this, remember to save your file after you have saved a
notebook and to re-initialize the DHCP server after changing or adding
options. You might start to wonder why your new options are not working.
Windows NT can provide the BOOTP client an address and a file name, but it
cannot provide the boot file. Neither NFS nor TFTP daemon is delivered with
Windows NT, although there are many shareware utilities; so our IBM
Network Station was not able to receive its boot file from Windows NT.
However, Windows NT can provide the IP address of a boot file server. We
set up the Windows NT DHCP server to provide the IP address of one of our
OS/2 DHCP clients as the boot file server. We configured the OS/2 client to
Chapter 2. Up and Running with DHCP
85
autostart the inet daemon, TFTP daemon (started from inetd), and the
portmapper. We rebooted the OS/2 DHCP client to restart everything. We
configured the NT DHCP server so that the boot file server provided the
BOOTP client (NC) with the address of the OS/2 client. After restarting the NT
DHCP server, we powered on the NC and it booted up from the OS/2
workstation. So, the Windows NT DHCP server provides a type of BOOTP
service. If you want more than just an address (for example, you actually want
to network boot a workstation), you will need to provide the name of a boot
file server, or add some third party software to Windows NT; so it can provide
TFTP services (we did not test any third party add-ons). Note that IBM has
Networkstation kits for different server platforms that provide file transfer
services.
2.5.6.2 DHCP
Next, configure the Network Station as a DHCP client. No changes are
required on the server to boot the Network Station as a DHCP client.
There is still one problem on the Network Station, which may be due to the
older firmware in our machine (Boot Monitor Version 3.0.1). On more than
one occasion, even though the Network Station was configured to get its IP
address from the network, it insisted on using the IP address that was
configured in NVRAM. We solved this problem by setting the NVRAM
addresses to all zeros, except for the subnet address.
2.5.7 AIX
Note
Before configuring AIX as a DHCP client, ensure that the file set
bos.net.tcp.client is installed. To verify this, use the command:
lslpp -l “bos.net.tcp.*”
To install this file set you will need AIX CD-ROM 1, and you may be
prompted for further CD-ROMs.
An AIX workstation can also be configured as a DHCP client, but then it
cannot be used as a DHCP server. We configured a DHCP client on AIX
Version 4.3.2 using the following procedure:
1. Open the Application Manager, and you will be presented with a window
as shown in Figure 60 on page 87.
86
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 60. (AIX Version 4.3.2) Application Manager
2. Double-click on the System_Admin icon, and you will open the System
Administration window (see Figure 61).
Figure 61. (AIX Version 4.3.2) Application Manager - System Admin
3. Double-click on the Network icon and you will be asked for the
administrator’s ID and password. After supplying these, the Network:
Loopback window will open (see Figure 62 on page 88).
Chapter 2. Up and Running with DHCP
87
Figure 62. (AIX Version 4.3.2) Network: loopback - Configuration
4. Double-click on the TCP/IP Configuration to edit the current TCP/IP
configuration.
5. Select the Basic Setup tab and then configure the Available network
interface and allow it to automatically obtain an IP address, (see Figure 63
on page 89).
88
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 63. (AIX Version 4.3.2) TCP/IP Properties: loopback - Configuration
6. We allowed our DHCP start options to take place now and at the system
restart. It is best to click on Apply now, and allow the system to apply any
changes. If TCP/IP is restarted successfully, then you will be returned to
this configuration screen and you may select OK.
Note
In order for DHCP (whether client or server) to function properly on an
RS/6000 machine, you must have DNS Services configured. This is due
to the fact that the Common Desktop needs name resolution working, or
else AIX is unable to start up TCP/IP correctly.
7. To check that TCP/IP is working properly, use the ping command.
We found that when using an AIX DHCP client on a network with DHCP
servers that were not OS/2 or AIX systems, the DHCP client would not
receive all of its DHCP options. The reason for this is that these other servers
do not assign all the configured options unless the client specifically requests
them. The work around for this is to add those options to the DHCPCD.INI file
on the AIX DHCP client.
Chapter 2. Up and Running with DHCP
89
We had to add the following lines to our DHCPCD.INI file in order to receive
Options one, six and three from a Windows NT DHCP server:
option
option
option
1
6
3
The complete file has been included on the CD-ROM that accompanies this
redbook.
2.5.8 Linux
Linux clients can certainly participate in a DHCP network, getting address
assignments from a DHCP server, such as OS/2 Warp Server for e-business.
Using instructions available on the Internet at:
http://www.redhat.com/corp/support/manuals/index.html
we tested RedHat Linux 6.0 with kernel 2.2.5-15. We dubbed this system
woody. Woody succeeded in picking up a dynamically assigned IP address
using either BOOTP or DHCP (both worked).
This version of Linux was far more stable than the previous one since we did
not have to apply a token-ring adapter patch for DHCP and we did not receive
any warning messages.
2.5.8.1 Enabling DHCP
To enable RedHat Linux’s DHCP, follow these steps:
1. Log on as user root.
2. Make sure that the user and group IDs for dhcpcd are both set to root, and
the permissions are set to read (r) and execute (x) for all users.
3. If it is not already running, start X-Windows with the command:
startx
4. To start the control panel, as shown in Figure 64 on page 91, open a
terminal and type the following command:
control-panel
90
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 64. [RedHat Linux] Control panel
5. Click on the network configuration icon (which is the icon on the right-hand
side of Figure 64).
Figure 65. [RedHat Linux] Interfaces section of network configuration
6. In the Interfaces section (see Figure 65 on page 91), edit tr0 (for token
ring) or eth0 (for Ethernet). Change the interface configuration protocol
(see Figure 66 on page 92) to DHCP (or BOOTP), then click on Done to
save these changes.
Chapter 2. Up and Running with DHCP
91
Figure 66. [RedHat Linux] Editing the tr0 (token-ring) interface
7. Click on Save, then Quit, to exit out of the network configuration section.
8. Shutdown and restart the Linux system using the command:
shutdown -h now
9. After rebooting, and after logging in, verify your address assignment by
using the command:
ifconfig
You can also verify network connections by using the ping command.
Linux is supported by its community of users and developers, whose input
can be found at:
http://www.redhat.com/community/user_groups.html
2.5.9 DOS, Windows 3.1, and Windows for Workgroups
In our test environment, we set up three DOS-based environments: DOS,
DOS with Windows 3.11, and DOS with Windows for Workgroups. The DOS
used was the IBM PC-DOS 2000, the only fully supported Year 2000 Ready
version of DOS.
All three environments, then, had the IBM DOS LAN Services (DLS) installed.
DLS is the DOS LAN client supplied with Warp Server Version 4, and it comes
with a TCP/IP stack from Network TeleSystems that supports DHCP.
Before installing DLS, make sure that you have all four diskettes and that you
are suitably licensed.
2.5.9.1 Enabling DHCP
To install DLS and enable TCP/IP, follow these steps:
1. Insert diskette 1, and enter the A:INSTALL command.
92
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
2. At the opening screen, press Enter to continue
3. When prompted to choose a directory, accept the default of C:\NET.
4. When prompted to enter the machine ID, enter the machine’s name. In this
instance, enter the NetBIOS name.
5. When prompted to enter the user name, enter the user ID of the person
who would normally use this machine.
6. When prompted to enter the domain name, enter the name of the domain
that the user would normally log onto.
7. The options panel will be displayed as shown in Figure 67.
Install for DOS LAN Services
============================
If all the options are correct, select 'The listed options
are correct,' and then press Enter. If you want to change
an option, use the Up Arrow or Down Arrow key to select it. Then
press Enter to see alternatives for that option.
+------------------------------------------------------------------------------+
| Graphical User Interface : Install GUI.
|
| Peer Services
: Install Peer Services.
|
| Windows Support
: Install Windows support.
|
| Protocol Driver
: IBM NetBEUI
|
|
|
| The listed options are correct.
|
+------------------------------------------------------------------------------+
Enter=Continue F1=Help F3=Exit
Figure 67. [DOS LAN Services] Installation options
8. If you do not require Peer support, scroll up to the Peer Services option,
and press Enter, then select the option Do not install Peer services.
9. If you do not require Windows support, select the Windows Support
option, then select the option Do not install Windows support.
10. Select the Protocol Driver option.
11. Select the Change driver for protocol option.
12. A panel will be displayed showing the protocols available for DLS, as
shown in Figure 68 on page 94.
If you are using native DOS, select TCPBEUI (Real-Mode) & IBM
NetBEUI.
Chapter 2. Up and Running with DHCP
93
If you are using Windows 3.11 or Windows for Workgroups, select
TCPBEUI (Windows Protect-Mode) & IBM NetBEUI.
We recommend dual protocols, as NetBEUI is still the most efficient within
a LAN. TCPBEUI is required to activate TCP/IP and can be used for WAN
(wide area network) communications.
Install for DOS LAN Services
============================
Please select the protocol used on your network.
+------------------------------------------------------------+
| Other protocol.
|
| Install 802.2 support
|
| TCPBEUI (Real-Mode)
|
| TCPBEUI (Real-Mode) & IBM NetBEUI
|
| TCPBEUI (Windows Protect-Mode)
|
| TCPBEUI (Windows Protect-Mode) & IBM NetBEUI
|
| IBM NetBEUI
|
+------------------------------------------------------------+
Figure 68. [DOS LAN Services] Selecting a protocol
13. When prompted for the IP address with the IPAddr= prompt, do not enter
an address, just press Enter.
A red error panel will be displayed (see Figure 69 on page 95), stating that
you have not specified a required value. Ignore the error message, and
press C to continue without setting the value.
94
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Install for DOS LAN Services
============================
The IP Address (IPAddr) is a required parameter for the
protocol you have chosen. This address is unique to your
computer. An example of an IP address would be:
IPAddr=123.2.43.7
+---------------------------------------------------------+
|
|t.)
| You did not specify the value of a required setting. |
|
|
|
To specify its value, press Enter.
|
| To continue Install without setting its value, press C. |
|
|
+---------------------------------------------------------+
IPAddr=________________________________________________________
Figure 69. [DOS LAN Services] Error when required value not specified
14. When prompted for the NetSubNetMask, GatewayAddr, and DNSAddr
parameters, do not enter values just press Enter. As before, press C at the
error panel to continue without entering a value.
15. Select Edit settings for protocol driver. A panel with the configuration
options will be displayed, as shown in Figure 70 on page 96.
Note
It is not readily apparent that the configurable options panel is only showing
a portion of the options. More options are visible if you scroll through the
list.
Chapter 2. Up and Running with DHCP
95
Install for DOS LAN Services
============================
The settings for your protocol driver are listed below. If
all the settings are correct, select 'The listed options
are correct'. Then press Enter. If you want to change a
setting, use the Up Arrow or Down Arrow key to select it.
Then press Enter to see alternatives for that setting.
Protocol Driver : TCPBEUI (Real-Mode) & IBM NetBEUI
+------------------------------------------------------------+
| ARP Cache Timeout in seconds=300
|
| DHCP Client ID=
|
| Virtual Circuits=16
|
| VC Receive Buffers=6
|
| VC Send Buffers=6
|
| Receive Window in Bytes=2920
|
| Send Window in Buffers=8
|
| Use Memory Option=UMB
|
|
|
| The listed options are correct
|
+------------------------------------------------------------+
F1=Help F3=Exit Esc=Previous Screen
Figure 70. [DOS LAN Services] Options for the protocol driver
16. Scroll up the list to the Bootp= option, and press Enter. Then select DHCP
from the list of options.
17. Select the DHCPClientID= option and enter the machine ID, as used
during step 4 on page 93. In this instance, we are entering the computer’s
TCP/IP hostname.
18. Press Enter on The listed options are correct, Driver configuration is
correct, and The listed options are correct options as they are displayed.
19. Select the Redirector.
For native DOS clients, select either the Use the basic, Use the full, or
Use the protect mode redirector options.
For Windows 3.11 or Windows for Workgroups clients, select Use the
virtual redirector.
20. Select the Startup option and then select Run DOS LAN Services only.
21. If necessary, select Network Card and choose the network adapter that
you use. See Figure 71 on page 97.
96
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Install for DOS LAN Services
============================
Select the type of network card that is installed in your
computer, and then press Enter.
+------------------------------------------------------------+
| IBM 16/4 Token Ring Credit Card Adapter
|
| IBM Token Ring (All Types)
|
| IBM Token Ring (MCA)
|
| IBM Token Ring II
|
| IBM Token Ring 4/16Mbs
|
| IBM Token Ring 4/16Mbs (MCA)
|
| IBM Token-Ring Network 16/4 Adapter II
|
| Intel EtherExpress 16 or 16TP
|
| Intel TokenExpress EISA 16/4
|
| Intel TokenExpress 16/4
|
| Novell/Anthem NE2000
|
| Novell/Anthem NE/2
|
| Olicom 16/4 Token-Ring Adapter
|
| Proteon Token Ring (P1392)
|
| SMC (WD) EtherCard PLUS/A (MCA) (WD 8003E/A or 8003ET/A) |
| SMC (WD) EtherCard PLUS Elite 16 Combo (WD/8013EW or 8013EW|
+------------------------------------------------------------+
Figure 71. [DOS LAN Services] Network adapter types
22. Select The listed options are correct.
You will be re-prompted to enter the IPAddr, NetSubNetMask,
GatewayAddr, and DNSAddr options. As before, leave these entries blank
and press C at the error panel to continue without entering a value.
23. Select The listed options are correct.
24. DLS will now be installed. Change diskettes as required.
25. At the completion of the installation, you will be prompted to either press
Enter to restart, or F3 to exit to DOS. Press F3, then use a text editor to
modify the \NET\NETWORK.INI file.
Chapter 2. Up and Running with DHCP
97
[network]
computername=BURGER
lanroot=C:\NET
autologon=no
autostart=netbeui full messenger
username=USERID
domain=ARMONK
lslogon=yes
reconnect=no
passwordcaching=no
timesync=yes
Figure 72. [DOS LAN Services] Extract from NETWORK.INI
Figure 72 shows an extract from the NETWORK.INI file, with modifications
that we have made to the defaults highlighted.
Note
The autostart parameters will vary depending on the Redirector option
chosen in step 19 on page 96.
26. Save the file, then restart the computer. As it reboots, you should see
messages similar to those in Figure 73 on page 99.
98
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
IBM Protocol Manager Version 2.1
IBM Token-Ring Network Driver Version2.7
NTSTS: Initialization started
Installing Network TeleSystems TCP/IOP
Version: 2.09 (960320)
Using VCs=16 VCSends=6 VCReceiveLarge=6
Driver NTSTS$ in section [NTS$NTST2] will bind with module [IBM$GENIBMTOK]
NTS$NTSTS2: Initialization succeeded
IBM DOS LAN Services Driver Version 5
Copyright (C) International Business Machines Corporation 1993, 1995
Copyright (C) Microsoft Corporation 1992
Transport Hooks Enabled
SHARE installed
IBM Netbind Version 2.1
Waiting for DHCP Server
IBM DOS NETBEUI 3.00
Copyright (C) International Business Machines Corporation 1993, 1995
Copyright (C) Microsoft Corporation 1992
The commmand completed successfully
C:\>
Figure 73. [DOS LAN Services] Messages displayed as a DOS workstation starts
27. You can now test your IP configuration, by using the PING command
(either in native DOS or by opening a DOS window from Windows). For
example, if we had used the TCP/IP name BURGER in step 17 on page
96, we would issue the command PING BURGER.
C:\>PING BURGER
PING - ICMP Echo Request/Reply 2.09 (960320).
Copyright (c) 1995 Network TeleSystems, Inc. All rights reserved.
PING burger (192.168.6.16): 56 ICMP data bytes
64 bytes from burger: icmp_seq = 0. time < 55 ms
----burger PING Statistics ---1 packets transmitted
1 packets received
0% packet loss
round-trip (ms) min/avg/max = 55/55/55
C:\>
Figure 74. [DOS LAN Services] Testing IP configuration using PING
2.5.10 Sun Microsystems Solaris
The Sun Microsystems SPARCstation 5 workstation is simple to set up as a
DHCP client. We wanted to use a token-ring card in our SPARCstation 5, but
Chapter 2. Up and Running with DHCP
99
we were unable to do so. We found a bug in the Sun code when using a
token-ring interface. Sun is aware of this and is working on it. If you really
need a token-ring adapter in your SPARCstation 5 workstation, you should
contact Sun. Our SPARCstation 5 workstation came pre-installed with SunOS
Version 5.6, which we upgraded to SunOS Version 5.7.
Before getting started on the client, we configured our DHCP server to serve
addresses for the 192.168.9.0 subnet. At the beginning of this chapter, we
showed you how to add a subnet to your DHCP server.
Since the token-ring card would not work with DHCP on the Sun machine, we
had to revert back to the Ethernet interface. To do this we executed the
following command:
/usr/sbin/ifconfig le0 plumb
We then configured a router between our Ethernet network (with the Sun
system) and our token-ring networks (the other systems), as shown in Figure
75 on page 100.
S un SPA RCstation5
192.168 9.22
E thernet
Adapter
192.168 6.4
Token-Ring
Adapter
192.168 9.0
E thernet
Subnet
192.168 9.4
E thernet
A dapter
192.168 6.0 Token-Ring
Subnet
Another
Router
B uffet
192.168.7.0
Token-Ring
Subnet
A IX-D HC P S erver for
subnets 192.168 6.0
and 192.168 9.0
Figure 75. Integrating the Sun Solaris workstation into our network
To verify that this configuration is functional, assign an IP address to the
interface with the command:
100
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/usr/sbin/ifconfig le0 192.168.9.145
and configured routing:
/usr/sbin/route add net 192.168.9 192.168.9.145
/usr/sbin/route add default 192.168.9.1
Then contact your DHCP server by executing:
/usr/sbin/ping 192.168.6.10
After verifying the router and network configuration, configure the
SPARCstation 5 workstation as a DHCP client. This is actually the easiest
machine to configure as a DHCP client. Simply create the following files:
• /etc/dhcp.le0
• /etc/hostname.le0
as shown in the following command:
/bin/touch /etc/dhcp.le0 /etc/hostname.le0
Now, you can simply reboot the UltraSparc workstation, or you can execute:
ifconfig le0 dhcp start
After a minute or so, our machine received the address 192.168.9.22, as
shown in Figure 76.
/usr/sbin/ifconfig le0
le0: flags=4863<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST,DHCP> mtu 1500
inet 192.168.9.22 netmask ffffff00 broadcast 192.168.9.255
ether 8:0:20:7e:38:27
Figure 76. DHCP results on the Sun SPARCstation 5
The configuration gets stored in the /etc/dhcp/le0.dhc file, a binary file.
You should also refer to your Sun documentation. The following man pages
may be useful as well:
•
•
•
•
dhcpagent (1M)
dhcpinfo
ifconfig (1M)
netstat (1M)
Chapter 2. Up and Running with DHCP
101
2.5.11 IBM WorkSpace On-Demand
Unlike other PC-based software, the IBM WorkSpace On-Demand does not
require any software to be loaded on a client system. All the software a PC
needs to operate is loaded from the WorkSpace On-Demand manager (based
on OS/2 Warp Server or OS/2 Warp Server for e-business) across the
network. All management, installation of new applications, software updates,
and other tasks are performed at the server. Consequently, WorkSpace
On-Demand costs a lot less to manage and support than other PC-based
solutions.
Because WorkSpace On-Demand supports modern Java and Web-based
applications, network managers need an easy way to set up TCP/IP for each
client. Managers can access WorkSpace On-Demand setup right from the
server or from a remote location across the network. Turning on DHCP for a
client merely involves selecting a button and clicking on Set. The WorkSpace
On-Demand manager can also act as the DHCP server, if needed.
Alternatively, a separate DHCP server (such as OS/2 Warp Server for
e-business, AIX, or Windows NT) may provide the address information to a
WorkSpace On-Demand client.
For Information about WorkSpace On-Demand, please refer to the redbook
WorkSpace On-Demand Handbook Release 2.0, SG24-5117.
2.5.11.1 Enabling DHCP
To activate DHCP for a WorkSpace On-Demand Version 2.0 client, follow
these steps:
1. Log on to the WorkSpace On-Demand manager as a system
administrator.
2. Open the LAN Server Administration program (you can also type NETGUI
at an OS/2 command prompt).
3. Open the Local Workstation icon for the WorkSpace server.
4. Open the Remote IPL Requesters folder.
102
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 77. [WorkSpace On-Demand] Contents of Remote IPL Requesters folder
5. If you need to create a new icon for a new WorkSpace On-Demand client,
drag the WorkSpace template to any empty part of the folder with the right
mouse button, (see Figure 77). Otherwise, open any existing WorkSpace
client icon.
6. Select the Identity tab.
7. Fill in the Machine ID and Description for the client.
8. Select the System tab.
9. Configure the Boot drive ID, the Operating System and the Swap file
location.
10.Select the Hardware tab.
11.Fill in the Network adapter address (MAC address), the Machine class
(you may select a generic type), the Network adapter, the Remote boot
method, the Video monitor and the Video resolution.
12.Select the Protocols tab.
13.Identify the Network protocol you wish to use. You may also wish to select
an Additional protocol to use, after the bootup process has completed.
14.Select the IP Address tab.
15.Make sure the Configure TCP/IP checkbox is selected.
16.Click on the Automatic, using DHCP radio button to turn DHCP on (see
Figure 78 on page 104). You may also select Also using DDNS (see
Chapter 4, “Serving names” on page 131) if you have a Dynamic DNS
server available on the network, such as Warp Server for e-business.
Chapter 2. Up and Running with DHCP
103
Figure 78. [WorkSpace On-Demand] IP address setup for client
17.After choosing any of the other settings required elsewhere in this
notebook, click on Create to save and close the settings for this client.
You may repeat these steps as often as needed to set up each WorkSpace
client. If you have many WorkSpace clients to set up (or modify), you may
automate the procedure with a REXX script.
We tested both the WorkSpace On-Demand clients and the WorkSpace
manager with DHCP servers running on OS/2 Warp Server for e-business,
AIX, and Windows NT. We found them to operate correctly with all these
DHCP servers. (If you opt to assign a dynamic IP address to the WorkSpace
On-Demand manager, we recommend that you also assign a fixed hostname
using a Dynamic DNS server. See Chapter 4, “Serving names” on page 131.)
2.5.11.2 Related publications
The IBM redbook titled IBM WorkSpace On-Demand Handbook Release 2.0,
SG24-5117 provides some additional details on setting up WorkSpace
On-Demand. In particular, Section 6.1.2 includes information on setting up a
104
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
large batch of WorkSpace clients using a REXX script and the NET
RIPLMACH command file. For Windows clients, there is the redbook,
WorkSpace On-Demand 2.0 Feature for Windows Clients, SG24-5396.
2.5.12 Hewlett-Packard LaserJet 4000 Printer (JetDirect)
We configured the HP 4000 Laserjet Printer with the JetDirect Token-Ring
card. The firmware on the JetDirect Token-Ring card we used was G.05.35.
Please note that there are other methods of configuring addresses on
JetDirect cards, which are described in the Hewlett-Packard manuals shipped
with the product. For example, you can use RARP. Please refer to your
manuals for further information beyond what is described here.
2.5.12.1 DHCP
We set up our DHCP server to reserve an IP address for our printer. If we
want to telnet to the printer to review or change some settings, it is quicker
and easier to do so if we already know the printer address. We reserved an
address on our DHCP server for the printer named potato, as shown in Figure
79 on page 106. We used a telnet session with our printer to configure it to
act as a DHCP client. To find out the initial IP address that the printer was
using, we used the buttons on the printer itself to print its configuration.
Chapter 2. Up and Running with DHCP
105
Figure 79. [Warp Server for e-business) DHCP printer configuration
Figure 80 on page 107 lists commands you can execute when you telnet into
your printer.
106
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 80. Executable commands at the HP printer through a telnet session
We executed these commands initially:
dhcp-config: 1
host-name: POTATO
We executed a telnet session with the printer to see its configuration after it
booted up from the DHCP server. This is shown in Figure 81 on page 108.
Chapter 2. Up and Running with DHCP
107
Figure 81. HP printer configuration after boot up
After configuring the DHCP server and the printer, you can use the printer
directly from clients that run a line printer daemon. You can also run a line
printer daemon on a Warp Server and share a printer resource with your LAN
clients. In this case, the line printer daemon would run on the server itself. To
share the printer with your LAN clients, follow these steps.
1. Open the TCP/IP Configuration GUI and select lprportd to start
automatically. Save and exit the window.
2. Start the line printer daemon manually before creating the printer object.
At an OS/2 command prompt, type the following two commands:
DETACH LPRPORTD
DETACH LPD
3. Create a printer object by dragging a printer off of the printer template.
4. Choose the correct printer driver (or install a new one) and also choose
one of the \PIPE\LPD ports, as shown in Figure 82 on page 109.
108
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 82. [OS/2 Warp] Create LPD printer
5. Right-click on your chosen \PIPE\LPD port after your have highlighted it.
Fill in the blanks, as shown in Figure 83 on page 110, using your printer
address for the LPD server. The LPD printer is either raw1 or text1 on a
single port JetDirect card. If you have a multi-port JetDirect box, the
hardware ports correspond to the number (for example, raw1, raw2, raw3
for a 3-port JetDirect box). Click Create, when you are done.
Chapter 2. Up and Running with DHCP
109
Figure 83. [OS/2 Warp] LPD Settings notebook
Your printer can now be shared via Warp Server like any other printer
resource, or you can print to it the UNIX way by using the LPR command.
If you want to print directly to the printer from the server or a client, you can
type the following command:
LPR -b -p RAW1 -s 192.168.6.11 <filename>
2.5.12.2 BOOTP
Officially, Hewlett-Packard does not support BOOTP on the JetDirect card,
unless you are using the HP JetAdmin software on the Windows 95, Windows
98, Windows NT, HP-UX, or Sun Solaris platforms.
For more information please consult the HP JetDirect Print Server Software
Installation Guide, which is available from:
http://www.hp.com/cposupport/manindex/jd600t_en_man.html
We were able to get the printer to work, but not completely as expected. We
could not get the printer to consistently accept the IP address that we wanted
110
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
it to use. We suggest that you use this device as a DHCP client rather than a
BOOTP client.
Chapter 2. Up and Running with DHCP
111
112
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 3. DHCP server and client interaction
To help provide a basis for going Beyond DHCP, this chapter takes a close
look at the function of the Dynamic Host Configuration Protocol. We start with
a high-level look at DHCP, options, and leases. Then, we illustrate the theory
presented in the RFCs for DHCP by looking at network traces of clients
interacting with servers at various stages of the configuration process.
The LAN traces discussed in this chapter were captured in our lab using the
Network Associates Sniffer-Basic product. If you’d like to follow along with
your own protocol analyzer, the trace files can be found in a variety of formats
on the CD-ROM that accompanies this redbook.
3.1 DHCP basics
As we discussed in 1.5.1, “BOOTP and the Dynamic Host Configuration
Protocol (DHCP)” on page 10, DHCP provides a standards-based method of
centrally managing and dynamically setting critical configuration parameters
for an IP host on a network. In a working DHCP environment, hosts may be
added to the network or moved between networks without the need for
manual configuration changes.
DHCP expands on existing BOOTP transport standards to achieve a method
of automatically passing a wide variety of configuration information, known as
options to clients. In addition to the IP interface address and subnet mask,
addresses for gateways, name servers, and many other items can be
managed and dynamically configured using DHCP servers. Appendix D,
“DHCP options (RFC 2132)” on page 457 has detailed information on the
broad scope of DHCP configuration.
Hosts that implement DHCP use the protocol to configure some or all of their
IP interfaces. Each interface is configured independently.
3.1.1 Client states
A DHCP client is in one of six states at all times. The states are defined by
RFC 2131 and determine how the client interacts with the DHCP server, and
how the server responds. The states are described in Table 4 below.
© Copyright IBM Corp. 1998, 2000
113
Table 4. DHCP client states
State
Description
INIT
The client has not had a lease and is
locating a DHCP server.
INIT-REBOOT
The client has had a lease and is
restarting its IP interface.
SELECTING
The client has received lease offers from
one or more DHCP servers.
BOUND
The client has been successfully
configured and has a valid lease.
RENEWING
The client is attempting to renew the lease
with the same DHCP server.
REBINDING
The client could not renew the lease with
the original server and is attempting find
another.
3.2 The DHCP lease
The configuration parameters managed by DHCP are valid for a period of
time called a lease. The duration of the lease is set when the client accepts a
configuration offer from a DHCP server. Although the protocol provides a
method for a client to request a specific length for the lease, there is actually
no negotiation of lease terms. The client either accepts or rejects the
parameters offered by a server.
3.2.1 Getting the lease
DHCP works with lease times using numbers of seconds as opposed to
specific dates and times. This avoids the need to synchronize clocks among
multiple hosts on a network.
When a client configures an IP interface using DHCP, three values related to
lease time are determined by the server and sent to the client using these
options:
114
Option 51
IP address lease time
Option 58
Renewal time value (T1)
Option 59
Rebinding time value (T2)
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The RFC for DHCP (RFC2131) suggests that the values for T1 and T2 default
to 50% and 87.5% of the lease duration, respectively. Some DHCP server
implementations only allow you to configure option 51 and derive the other
values according to the suggested defaults. On others servers, such as AIX,
OS/2 and Shadow IPserver, you are able to specify all three options. The
RFC mandates that the lease time must be greater than T2, and that T2 must
be greater than T1.
Although a client’s DHCP implementation might use these values to calculate
specific dates and/or times for lease events to occur, conceptually it’s easier
to think of them as simple counters. The counters are set to the specified
number of seconds at the beginning of the lease, and decrement at the rate of
one per second.
3.2.2 Renewing the lease
When the T1 counter reaches zero, the client’s state changes from BOUND to
RENEWING. At this time the client attempts to renew its lease with the same
DHCP server that originally provided the configuration. If it is successful, the
counters are reset to freshly calculated values. The client now returns to the
BOUND state.
If the client is not able to contact the DHCP server, it continues to try
periodically. While the client is in the RENEWING state, no disruption of IP
traffic through the interface occurs.
3.2.3 Rebinding the lease
Since T2 must be greater than T1, the T2 counter only expires if the client
was not able to renew the lease with the original DHCP server. As a result of
this failure, the client’s state changes from RENEWING to REBINDING.
At this point, the client tries to get a lease from any DHCP server. If the
operation is successful, the counters are set to the values specified in the
new lease. The client then returns to the BOUND state.
If the client is unable to contact any DHCP server, it continues to try for the
remainder of the current lease. As with the RENEWING state, a REBINDING
client is still able to transmit and receive using its assigned IP address with no
interruption. The change in state from BOUND to RENEWING to REBINDING
is transparent to the applications that are using the IP interface.
Chapter 3. DHCP server and client interaction
115
3.2.4 Eviction
While you won’t find the term eviction in any DHCP RFC, it does accurately
describe what happens to a host’s IP interface if the lease counter reaches
zero. At this point, the lease has expired and the interface can not transmit or
receive datagrams using the previously assigned IP address. Any
applications on the host that had been using the interface receive errors.
It is important to note that eviction only occurs when a client has not been
able to contact any DHCP server on the network and has made multiple
attempts over a considerable amount of time.
To recover from this undesirable situation, the client changes to the INIT state
where it configures the interface with an address of 0.0.0.0, and continues its
attempts to locate a DHCP server.
3.3 Interaction between DHCP server and client
The sections that follow provide more detail on the information that is
exchanged between the DHCP server and client during the client boot and
TCP/IP startup process.
Figure 84 on page 117 shows an example of DHCP client and server
interaction from the viewpoint of a network trace. In all, four messages are
exchanged between the client and server, and result in the successful
configuration of the client’s interface. We’ll take a closer look and the content
of these messages as the chapter continues.
116
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 84. [Sniffer] Trace of DHCP interaction
3.3.1 BOOTP
The boot protocol (BOOTP) provides a method for datagrams to be sent and
received on an interface prior to the assignment of an IP address. This is very
handy, since a goal of DHCP is to assign this address. Clients use an IP
address of 0.0.0.0 during their initial interactions with the DHCP server. This
is shown in the IP decode portion of Figure 84 on page 117. Also note that
frames from the client are broadcast at both the network layer
(255.255.255.255) and the physical layer. The Sniffer uses the word
Chapter 3. DHCP server and client interaction
117
Broadcast in the Destination Address column to indicate a physical layer
address of xFFFFFFFFFFFF.
DHCP uses the User Datagram Protocol (UDP) to carry traffic between
clients and server. In the UDP decode section of Figure 84, we see that well
known ports 67 and 68 are used for DHCP.
3.3.2 Relay agents
Broadcast traffic can be managed by network infrastructure, and is
sometimes limited to the logical or physical portion where it originates. By
taking advantage of relay agents, DHCP clients and servers can be located
virtually anywhere on the network.
Relay agents forward DHCP traffic to other subnets, and may be configured
with specific DHCP server addresses. This limits a DHCP client’s broadcast
traffic to the subnet of origin. Many routers can be configured to provide the
relay agent function. These agents may be called BOOTP relay agents or
DHCP relay agents. OS/2 Warp Server for e-business and Windows NT
servers include such agents.
3.4 DHCP messages
The basic unit of communication between the client and server is called a
DHCP message. Configuration parameters (options) are carried in the body
of the message. Each of the eight DHCP message types is described below.
3.4.1 DHCPDISCOVER
This message is used when the client is configuring its interface using DHCP
for the first time, or when it is has been unable to reuse a previously assigned
address. At this point the client is said to be in the INIT state. Figure 85 on
page 119 takes another look at frame #7 in the trace, this time decoding the
DHCP portion of the data.
118
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 85. [Sniffer] DHCPDISCOVER message
The client broadcasts the DHCPDISCOVER message to solicit offers from all
reachable DHCP servers on the network. Items of interest within the message
include the client’s MAC address, its host name and the IP address it last
used (if any). In the case depicted in Figure 85, our client (bob765l) had
previously used 9.3.240.169 for its IP address.
At the bottom of the screen we see the client requesting eight parameters.
The numbers to the left of each parameter are the DHCP option numbers.
Chapter 3. DHCP server and client interaction
119
At this point, the client expects to get a DHCPOFFER message from one or
more servers. If the client does not receive a response to its request, it waits
a period of seconds before re-sending the DHCPDISCOVER message.
3.4.2 DHCPOFFER
A server sends the DHCPOFFER message in response to the
DHCPDISCOVER message from the client. If the server has no suitable IP
address available, it will not respond to the client’s request. The IP address,
mask, domain name, lease intervals, and other options are contained in this
datagram. DHCP servers may temporarily reserve any offered IP addresses,
so they will not be offered to several DHCP clients at the same time.
Figure 86 on page 121 shows the decode of the DHCPOFFER message sent
by the server in response to our client’s request. At the top of the image is a
column titled Delta Time, indicating about two seconds elapsed before the
server responded to the client’s DHCPDISCOVER frame. During this time, the
server was checking that the IP address it would offer was not already in use
on the network. The server’s ARP activity was filtered out of the trace for the
purpose of clarity.
Note that the Transaction ID value given in Figure 84 on page 117 and Figure
85 on page 119 are the same. This ID mechanism is used to give the client
the ability to match replies to a specific request message it sent.
120
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 86. [Sniffer] DHCPOFFER message
In the lower portion of the decode, we see the various configuration
parameters being offered by this server. The duration of the lease along with
the renewing and rebinding intervals are shown at the bottom of the list of
parameters.
Upon receipt of a DHCPOFFER message from one or more servers, the client
is said to be in the SELECTING state. Appropriately named, as the client will
compare offers from each server, and select the one it deems most suitable.
The criteria for this selection as well as time allowed for selecting vary by
platform.
Chapter 3. DHCP server and client interaction
121
3.4.3 DHCPREQUEST
After selecting the DHCPOFFER message to work with, the client sends a
DHCPREQUEST message that includes a Server Identifier (an IP address) in
the body of the message. The Server Identifier is the IP address of the DHCP
server whose offer is being accepted. The DHCPREQUEST message is sent
as a broadcast datagram. This is done so that any other server that had sent
an offer would also receive the message. In this manner, a client accepts
parameters from one server and rejects other offers with a single datagram.
The body of the DHPCREQUEST message contains the same parameters
the sever sent in its DHCPOFFER frame, and adds a field called Client
Identifier. This identifier is important during lease renewal or after a client
reboot.
The content and addressing of the DHCPREQUEST message vary slightly
based on the state of the client. A discussion of the other four types of
DHCPREQUEST messages and how each relates to the state of the client
follows.
3.4.3.1 DHCPREQUEST from the INIT state
A client is said to be in the INIT state when it is using DHCP for the first time.
It also enters the INIT state when it is not able to reuse a previous DHCP
assigned configuration.
While in the INIT state the client must use 0.0.0.0 as its IP address. This is
shown in Figure 87 on page 123.
122
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 87. [Sniffer] DHCPREQUEST message from a client in INIT state
3.4.3.2 DHCPREQUEST from the RENEWING state
At the point when the client’s T1 timer has expired, the client leaves the
BOUND state and enters the RENEWING state. At this time, the client
attempts to renew his or her lease with the same server that originally
assigned his or her address. This DHCPREQUEST message is addressed
directly to the leasing server, who responds with a DHCPACK message. After
receiving this acknowledgment from the server, the lease timers are reset,
and the client returns to the BOUND state.
Chapter 3. DHCP server and client interaction
123
3.4.3.3 DHCPREQUEST from the REBINDING state
If the client’s T2 counter expires, the end of the lease is approaching. This is
called the REBINDING state. In this state, the DHCPREQUEST does not
contain a server identifier, and the datagram is broadcast. A client enters the
REBINDING state only after attempting to renew the lease with the original
server (RENEWING state).
If a client is not successful in renewing or rebinding the lease on its
configuration, it must stop using the current configuration information and
enter the INIT state.
3.4.3.4 DHCPREQUEST from the INIT-REBOOT state
If a client previously had successful configuration via DHCP and is restarting
the IP interface, it is said to be in the INIT-REBOOT state. The client
configures its IP interface with the address it last used, and broadcasts a
DHCPREQUEST message. The server checks that the client identifier value
contained in the body of the message is the same as when the address was
previously assigned before responding with a DHCPACK message. Use of the
INIT-REBOOT state can help a client retain the IP address, which is desirable
in many cases.
Figure 88 on page 125 shows a DHCPREQUEST message from a client in
the INIT-REBOOT state.
124
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 88. [Sniffer] DHCPREQUEST message from a client in INIT-REBOOT state
Use of the INIT-REBOOT state can save time because the client neither
broadcasts the DHCPDISCOVER message nor compares offers, and the
server does not test the requested IP address.
However, there are situations where the attempt to quickly re-use the prior
configuration can actually take more time. We created such a situation in our
lab to illustrate this scenario.
At the top of Figure 88, note that the client actually transmitted the
DHCPREQUEST message five times over a 15 minute period and did not
receive a response. The client eventually changes from the INIT-REBOOT to
Chapter 3. DHCP server and client interaction
125
the INIT state and starts the configuration over with the DHCPDISCOVER
message in frame #7. The time required for the client to get back to the INIT
state can be virtually eliminated if a DHCP server is able to send it a
DHCPNAK message. This type of message is discussed in the next section.
3.4.4 DHCPNAK
A DHCP server sends this response to the DHCPREQUEST message only if
it can determine that the parameters being requested by the client are not
appropriate for the network. When a client receives a DHCPNAK message, it
immediately returns to the INIT state and restarts the configuration process
with the DHCPDISCOVER message.
To create the scenario for Figure 88 on page 125 (where the client stayed in
INIT-REBOOT for 15 minutes), we moved the client from a network that was
completely unknown to our DHCP server. Since the server had no
authoritative knowledge of the clients prior network, it did not respond with a
DHCPNAK.
When a server is controlling address assignment for multiple subnets, it can
speed up the retry process by sending a DHCPNAK message. This is
illustrated in Figure 89 on page 127, where we moved a client from the
192.168.8.0 network to the 192.168.6.0 network. In this case, both subnets
are being managed by the same DHCP server.
126
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 89. [Sniffer] DHCPREQUEST followed by a DHCPNAK
In frame #23 the client requests a specific IP address that is not valid for the
network it is on. In frame #24 the server responds with a DHCPNAK
message. The client returns to the INIT state at once, and sends the
DHCPDISCOVER message (frame #25).
3.4.5 DHCPACK
This message is similar in content to the DHCPOFFER. The server sends this
message in response to the DHCPREQUEST from the client as verification
Chapter 3. DHCP server and client interaction
127
that the client owns the IP address. If the client was renewing the lease, the
DHCPACK resets the lease timers and the client returns to the BOUND state.
If the client is in the INIT or INIT-REBOOT state, the configuration of a client’s
interface is not complete until this message is received. After receipt of this
message the client should also test (via ARP) that the address is not in use. If
the ARP test is successful, configuration is complete and the client is in the
BOUND state.
3.4.6 DHCPDECLINE
This message is broadcast by the client to indicate that an IP address that
had been requested, offered, and acknowledged was found to be in use when
the client tested it. The DHCP server will note that the address is in use and
will not offer it to other clients.
3.4.7 DHCPRELEASE
This message is used by clients to formally cancel their lease. Most clients do
not send this message during a shutdown so that they may attempt to use the
same IP address when they are rebooted.
The DHCPRELEASE message is a sent within a unicast datagram to the
DHCP server that holds the lease on the client’s configuration. An example of
the DHCPRELEASE message is shown in Figure 90 on page 129.
128
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 90. [Sniffer] DHCPRELEASE message
Upon sending the DHCPRELEASE message, the client’s interface returns to
the INIT state, and can no longer transmit or receive datagrams using its
former IP address. For this reason the server does not acknowledge receipt
of the DHCPRELEASE message. It does of course, mark the address as
unused.
3.4.8 DHCPINFORM
A client that has a manually configured IP address may use this message to
request additional configuration parameters from the DHCP server. This
allows hosts who must have a manually defined interface address to take
advantage of DHCP’s ability to deliver other configuration options. We did not
take a Sniffer trace of DHCPINFORM.
Chapter 3. DHCP server and client interaction
129
130
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 4. Serving names
This chapter describes how to assign unique names, statically and
dynamically, to all devices on your TCP/IP network. We will see how a
dynamic name server works and how it integrates with existing static name
servers and firewalls. For information regarding justification of a DNS scheme
and how to design a domain name schematic, please refer to 1.6, “Domain
Name Services (DNS)” on page 12.
4.1 Name serving on the Windows platforms
The following sections will cover Windows 95/98/NT in various static and
dynamic DNS scenarios. We first cover Windows NT as a static DNS server,
as this will lead into the configuration of Windows NT as a DDNS server using
WINS (Windows Internet Naming Service). After we have given you the
server configuration scenarios for DNS and DDNS on Windows NT, we will
cover the Windows 95/98/NT client scenarios with DNS and DDNS.
4.1.1 Windows NT as a Static DNS server
This section shows you how to set up your Windows NT server as a static
DNS server using an example of a simple internal network environment. Note
that Windows NT’s implementation of DNS cannot be configured as a
dynamic DNS server.
4.1.1.1 Scenario
In this section, we use a simple network environment with only one subnet.
This network is not connected to the Internet or any other networks. In this
network, we assume that we have a DHCP server on the Windows NT server
up and running. We also have DHCP client computers. In this network we
initially do not have a DNS server.
We configure the Windows NT server as a static DNS server. Later in this
chapter, we will also configure Windows NT server as a WINS server.
In our environment, our domain name is bellevue.cooking.net and the
DHCP/DNS server’s hostname is mustard. Table 5 illustrates our network
configuration:
Table 5. Windows NT server configuration information
Configuration settings
Value
Network ID
192.168.8.0
© Copyright IBM Corp. 1998, 2000
131
Configuration settings
Value
Subnetmask
255.255.255.0
DHCP/DNS/WINS Server IP address
192.168.8.10
DHCP/DNS/WINS Server Hostname
mustard
4.1.1.2 Tasks
The tasks to complete the scenario are as follows and the sections following
this describe the steps of each task:
1. Plan the primary domain.
2. Install the DNS server service on the Windows NT server.
3. Configure DNS search order on the Windows NT machine.
4. Create a DNS primary name server using the DNS manager.
5. Create a new zone for the forward mapping.
6. Create a new zone for the reverse mapping.
7. Configure hosts.
8. Add DHCP options for a DNS server.
9. Enable WINS lookup, if desired, and add DHCP options for the WINS
server.
10. Start the DNS server.
4.1.1.3 Planning
The first step is to design your domain. Before you establish domain name
servers in your network, you need to determine a domain name for your
domain. You also need to determine the server that is going to be the primary
DNS server for your domain. In our scenario bellevue.cooking.net is the
primary domain on the Mustard name server. All the hosts on the 192.168.8.0
network are included in the domain bellevue.cooking.net.
4.1.1.4 Setting up
The default installation of Windows NT Server 4.0 does not install the DNS
server service. To install the DNS server service, follow the steps below:
1. Open the Control Panel.
2. Double-click the Network icon.
3. On the Network notebook, click on the Services tab and click on the Add
button.
132
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
4. Select Microsoft DNS Server from the list and then click OK.
5. When prompted, enter the path to the Windows NT files (Windows NT 4.0
Server CD) and click Continue to copy the files and finish the installation.
6. After installation, restart your computer.
7. Upon restart you will have to reapply any fixpacks if any were installed
before the DNS services installation.
4.1.1.5 Verifying TCP/IP configuration
The following steps describe how to set up your Windows NT computer to run
the DNS server. All the values we used to configure in the following steps are
the values we defined in our sample network.
Before you start configuring, make sure that TCP/IP is configured correctly.
The DNS server must have a static IP address and, therefore, cannot be a
DHCP client.
The DNS server computer itself can also be a DNS client. Follow the steps
below to add the DNS server information:
1. Open the Control Panel and then open the Network icon.
2. Click on the Protocols tab.
3. Select TCP/IP Protocol and click on Properties.
4. Click on the DNS tab and type a hostname and domain name.
5. Under DNS Service Search Order click Add.
6. In the DNS server box type the IP address for your computer (DNS server)
and click Add.
7. Click OK to close the notebook.
8. Click OK to close the Network notebook.
4.1.1.6 Configuring DNS server
The following steps illustrate how to set up a static DNS server:
1. By default, the Microsoft DNS server service starts automatically when the
computer is started. If the DNS server is not running, open the Service
icon in the Control Panel, select the Microsoft DNS Server in the service
list, and click on the Start button. You can also start the DNS server at the
command prompt by using the following command:
NET START DNS
2. Open the DNS manager. To do so, select [Start => Programs =>
Administrative Tools (Common) => DNS Manager].
Chapter 4. Serving names
133
3. Click the DNS menu and select New Server.
4. Enter the hostname of the DNS server and click OK.
Figure 91. [Windows NT] Adding DNS Server
5. Next, you need to add a new zone for the forward mappings. In the Server
list highlight the DNS server you have just created and select New Zone
from the DNS menu.
6. Check Primary for the zone type and click Next>.
7. Enter your full domain name for the zone name. Then, move your cursor to
the space for the file name. <ZONE_NAME>. DNS is the default name for
the zone file. If you want a different name, change the file name and click
Next>.
Figure 92. [Windows NT] Creating a new zone for forward mappings
8. Click Finish to create the zone.
9. If you want to have reverse lookups you need to create a zone for the
reverse mappings. Repeat steps five through eight above. Naming is done
differently for reverse mapping. For example, the IP address for mustard is
192.168.8.10, making the zone name for the reverse domain
134
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
8.168.192.in-addr.arpa. If our IP address was 9.19.24.10, the zone name
for the reverse domain would be 24.19.9.in-addr.arpa.
Figure 93. [Windows NT] Creating a new zone for reverse mappings
10. Select Update Server Data Files from the DNS menu to affect the
changes immediately to the zone files. It may also be necessary to stop
and then start the DNS services through the Services icon in the Control
Panel.
4.1.1.7 Adding hosts
The next step is to add hosts to the domain.
1. Highlight the zone for the forward mappings and select New Host from the
DNS menu.
2. Enter the hostname and IP address. If you made a zone for reverse
mappings, check Create Associated PTR Record then click Add Host.
Figure 94. [Windows NT] Adding a New Host
3. Repeat steps one and two above for all the hosts in your network.
4. When you finish adding hosts, click Done.
Chapter 4. Serving names
135
4.1.1.8 Adding DHCP options for DNS server
You need to configure the DHCP server so that DHCP clients can get the
configuration information for the DNS server and domain name. In our
scenario, we added following two options:
• Option 006 DNS servers: IP address of the DNS server
• Option 015 Domain name: Domain name
Refer to 2.2.3, “Basic configuration” on page 35 to see how to add DHCP
options.
4.1.2 Windows NT as a DDNS using WINS
In 4.1.1.5, “Verifying TCP/IP configuration” on page 133, we discussed how to
set up a static DNS server. Assuming you now have the DHCP server and the
static DNS server running in your network, what happens to name resolution
when a host releases its IP address and obtains a different IP address next
time? Since DNS is a static database, the administrator must update it
manually.
By integrating the static DNS server with the WINS server, you can resolve
the names dynamically by having the DNS server update the A record for the
DHCP client.
WINS (Windows Internet Name Service) servers maintain a database that
maps NetBIOS computer names to TCP/IP addresses. WINS supports
dynamic computer name registration and name resolution. Each time a WINS
client starts, it registers its NetBIOS name - IP address mappings with the
configured WINS server. Although WINS provides dynamic name services, it
offers a NetBIOS name space; so you need to integrate WINS server with the
DNS server to resolve TCP/IP hostnames dynamically. This is achieved by
having a DNS server point to a WINS server for both forward and reverse
name resolutions. By enabling WINS lookup, if a mapping is not found in the
DNS database, DNS server queries a WINS database.
Figure 95 on page 137 shows the interaction between a DHCP server, WINS
server, and a static DNS server.
136
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 95. [Windows NT] DHCP – WINS – DNS interaction
4.1.2.1 Installing WINS server
If WINS server is not installed, follow the steps below:
1. Open the Control Panel.
2. Open the Network folder.
3. On the Network notebook click on the Services tab and click on the Add...
button.
4. Select Windows Internet Name Server from the list and then click OK.
5. When prompted, enter the path to the Windows NT files (Windows NT 4.0
Server CD) and click on Continue to copy the files and finish the
installation.
4.1.2.2 Configuring WINS server
By default, the WINS server service starts automatically when the computer
is started. If the WINS server is not running, open the Service icon in the
Control Panel, select Windows Internet Service in the service list and click
on the Start button. You can also start the WINS server at the command
prompt by using the following command:
NET START WINS
Chapter 4. Serving names
137
To start the WINS Manager, select Programs from the Start menu, then
select Administrative Tools (Common). Then click WINS Manager. Click on
Server, then Add WINS Server and you will get a screen as seen in
Appendix 96, “[Windows NT] Adding a WINS server” on page 138. In that
window type in the IP address of your DNS, assuming WINS and DNS
services are going to be on the same address. WINS Manager can show
some basic statistics for the selected server. To display additional statistics,
click Detailed Information on the Server menu. You can also view all the
mappings by selecting Show Database from the Mappings menu.
Figure 96. [Windows NT] Adding a WINS server
To integrate the DNS server with a WINS server, you need to configure your
DNS server to enable WINS lookup for both forward and reverse mappings.
Follow the steps below:
1. Open the DNS Manager.
2. Highlight the zone (domain) for forward mappings. Select Properties from
the DNS menu.
3. Click on the WINS Lookup tab. Check Use WINS Resolution and enter
WINS server’s IP address. Click OK.
138
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 97. [Windows NT] Enabling WINS lookup
4. Highlight the appropriate in-addr.arpa zone (domain) for reverse
mappings. Select Properties from the DNS menu.
5. Click the WINS Reverse Lookup tab. Check Use WINS Reverse Lookup
and enter the DNS host domain to be appended to the NetBIOS name
before returning the response. Click OK.
Figure 98. [Windows NT] Enabling WINS reverse lookup
Chapter 4. Serving names
139
6. You also need to add some DHCP options. To add the options, see 2.3.3,
“Basic configuration” on page 45. Add and configure the following options
using the DHCP manager:
Option 044 WINS/NBNS Servers IP address of the WINS server
Option 046 WINS/NBT Node
0x8 (H-node)
Option 044 is used to configure the IP address of the WINS server. Option
046 is used to configure clients to use H-node. For more information about
H-node, refer to 5.2, “Resolving NetBIOS names to IP addresses” on page
232.
7. Restart the DHCP, DDNS, and WINS servers.
4.1.2.3 Configuration files
Table 6 shows DNS configuration files for the Windows NT server. After
configuring a DNS server, these files are created and stored in the
\WINNT\SYSTEM32\DNS directory.
Note
\WINNT is the default directory where Windows NT is installed.
Table 6. DNS configuration files
Name of the file
Description of the file
<zone_name>.dns
Database file
Contains resource records for the zone
Used for forward mappings
w.x.y.z.in-addr.arpa.dns
Reverse lookup file
Maps IP addresses to hostnames
cache.dns
Contains names and addresses of the root domain name
servers
Boot file
Used by manual start-up method *
* By default, the N T DN S server is configured to boot from the data in the registry.
Figure 99 on page 141 shows an example of a <zone_name>.dns file.
140
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
;
;
;
;
@
;
;
;
Database file bellevue.cooking.net.dns for bellevue.cooking.net zone.
Zone version: 4l
IN
SOA
mustard.bellevue.cooking.net.
Administrator.bellevue.cooking.net.
(
4
; serial number
3600
; refresh
600
; retry
86400
; expire
3600
)
; minimum TTL
Zone NS records
@
IN
NS
mustard
;
;WINS lookup record
;
@
0
IN
WINS
192.168.8.10
IN
A
192.168.8.10
;
;Zone records
;
mustard
Figure 99. [Windows NT] BELLEVUE.COOKING.NET.DNS file
Figure 100 on page 142 shows an example of a 8.168.192.in-addr.arpa file.
Chapter 4. Serving names
141
;
;
;
;
@
;
;
;
Database file 8.168.192.in-addr.arpa.dns for 8.168.192.in-addr.arpa zone.
Zone version: 4l
IN
SOA
mustard.bellevue.cooking.net.
Administrator.bellevue.cooking.net.(
4
; serial number
3600
; refresh
600
; retry
86400
; expire
3600
)
; minimum TTL
Zone NS records
@
;
;
;
IN
mustard.bellevue.cooking.net.
IN
WINS-R
bellevue.cooking.net.
IN
PTR
mustard.bellevue.cooking.net.
NBSTAT lookup record
@
;
;
;
NS
0
Zone records
10
Figure 100. [Windows NT] 8.168.192.IN.ADDR.ARPA.DNS file
4.1.3 Windows 95/98 as a static DNS client
To enable DNS client function in Windows 95/98, follow the steps below:
1. Open the Control Panel.
2. Double-click on the Network icon.
3. Highlight the TCP/IP protocol and click on the Properties button.
4. Click the DNS Configuration tab.
5. Check Enable DNS.
6. Enter a hostname and domain name. Enter IP address(es) of the DNS
server(s) and click Add.
4.1.4 Windows NT Workstation 4.0 as a static DNS client
To enable the DNS client function for Windows NT Workstation 4.0, follow the
steps below:
1. Open the Control Panel.
142
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
2. Double-click on the Network icon.
3. Click the Protocols tab, highlight TCP/IP Protocol in the Network
Protocol list and click on the Properties button.
4. Click on the DNS tab. Enter a hostname and domain name. Click Add. In
the DNS server box, enter IP address(es) of a DNS server(s) and click
Add.
4.1.5 Windows with IBM Dynamic IP Client
The Dynamic IP Client for Windows 95/98/ NT allows a user to have a
hostname mapped to their dynamically allocated IP address given by a DHCP
server. This is used in conjunction with the Microsoft DHCP client and is not
necessary for WINS services. However, it is necessary to implement a
Windows DDNS client with a non-Microsoft DDNS server. The software can
be found at:
http://techsupport.services.ibm.com/asd-bin/doc/en_us/win95cl/f-feat.htm
To use the IBM Dynamic Client for Windows 95/98/NT the following items are
required:
• Microsoft Windows 95 or Windows NT (Version 4.0 and higher) configured
with the TCP/IP protocol
• IBM DDNS server for OS/2 or AIX
• DHCP server (Clients must obtain its IP address using DHCP)
The following sections show you how to set up IBM Dynamic IP Client for
Windows NT and Windows 95.
4.1.5.1 Windows NT
To install and run the IBM Dynamic IP Client, the user logged in must be a
user who belongs to the Administrator Group. This is because only the
administrator may manually request/renew the lease or set/configure DHCP
information.
Before you configure DDNS client, make sure that DHCP is enabled on that
computer. See 2.5.3, “Windows NT Workstation 4.0” on page 73 for more
information.
To install the IBM Dynamic IP client, follow the steps below:
1. Open a command prompt. Change to the directory where IBMDYNIP.EXE
is located, and run the program. Or, just double-click on IBMDYNIP.EXE
from the Windows NT Explorer.
Chapter 4. Serving names
143
2. Follow the instructions on each panel to finish installation. When the
installation is finished, a message saying Installation is finished. The
IBM Dynamic IP Client Configuration Process will begin appears. Click
OK.
3. You can configure your DDNS client at this time or later. If you already
know all the information, type the hostname, domain name, and primary
domain name server (or just verify the current configuration) and click OK.
If you wish to configure it later, click Cancel.
4. Restart your computer.
If you did not configure the DDNS client when installing IBM IP Dynamic
Client, or if you want to change the configuration, follow the steps below:
1. Select [Programs — IBM Dynamic IP Client — Dynamic IP
Configuration].
2. Type your hostname, domain name, and primary domain name server and
then click OK.
Figure 101. [Windows NT] Dynamic IP client configuration
3. The DDNS update runs automatically. If the hostname update was
successful, an information window prompts you that the update was
successful.
4. To check the return code of the NSUPDATE, open the Dynamic IP
Configuration and click on the Diagnostic tab, as shown in Figure 102 on
page 145.
144
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 102. [Windows NT] Dynamic IP client Diagnostics tab
5. To check the current configuration, type IPCONFIG /ALL at the command
prompt.
DDNS updates occur automatically when a user manually renews or releases
the lease, using IPCONFIG.EXE, or when a user changes the client’s hostname,
using the Dynamic Configuration program.
4.1.5.2 Windows 95/98
Before you install and configure the DDNS client, make sure that DHCP is
enabled on that computer. See 2.5.2, “Windows 95 and Windows 98” on page
72 for how to configure DHCP client.
To install the IBM Dynamic IP Client, follow the steps below:
1. Open a command prompt. Change to the directory where IBMDYNIP.EXE
is located and run the program. Or, just double-click on IBMDYNIP.EXE
from the Windows Explorer.
2. Follow the instructions on each panel to finish installation. When the
installation is finished, a message saying Installation is finished. The
IBM Dynamic IP Client Configuration Process will begin appears. Click
OK.
3. You can configure your DDNS client at this time or later. If you already
know all the information, type the hostname, domain name and primary
Chapter 4. Serving names
145
domain name server and click OK. If you wish to configure it later, click
Cancel.
4. Restart your computer.
If you did not configure the DDNS client when you installed the IBM IP
Dynamic Client, or if you want to make changes to the configuration, follow
the steps below:
1. Select [Programs — IBM Dynamic IP Client — Dynamic IP
Configuration].
2. Type your hostname, domain name, and primary domain name server.
Then click OK.
Figure 103. [Windows 95] Dynamic IP client configuration
3. Reboot your computer.
4. To check the return code of the NSUPDATE, open the Dynamic IP
Configuration and click on the Diagnostic tab.
146
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 104. [Windows 95] Dynamic IP client diagnostics tab
5. To check the current configuration, type WINIPCFG at a command prompt.
When running on Windows 95, DDNS updates only occur automatically at the
correct DHCP renewal intervals. Manual renewal and release of the lease
using WINIPCFG.EXE does not coincide with automatic DDNS updates.
When running on Windows 95, you must reboot after changing the
configuration of the IBM Dynamic IP Client, since the configuration changes
and updates will not occur automatically.
4.1.5.3 Windows 95/98 using presecured mode
This section explains how to configure an IBM Dynamic IP Client on Windows
95 in the Dynamic Presecured Domain environment.
Before you install and configure a DDNS client, make sure that DHCP is
enabled on that computer. See 2.5.2, “Windows 95 and Windows 98” on page
72 for a description of how to configure a DHCP client.
To install the IBM Dynamic IP Client, follow the steps below:
1. Open a command prompt. Change to the directory where IBMDYNIP.EXE
is located and run the program. Or, just double-click on IBMDYNIP.EXE
from the Windows Explorer.
2. Follow the instructions on each panel to finish installation. When the
installation is finished, a window appears that says Installation is
Chapter 4. Serving names
147
finished. The IBM Dynamic IP Client Configuration Process will begin.
Click OK.
3. You can configure your DDNS client now or later. If you already know all
required information, type the hostname, domain name, and primary
domain name server and click OK. If you wish to configure at a later time,
click Cancel. Do not restart your computer at this time.
4. Copy a key file (<host_name>.key is the default name) in the
\WINDOWS\DDNSETC directory (\WINDOWS is the directory where
Windows 95 is installed) and rename the file name to DDNS.DAT. If you
already have a DDNS.DAT file, you can either replace or append the file to
the existing file.
5. Restart your computer.
If you did not configure the DDNS client when you installed the IBM IP
Dynamic Client, or if you want to make changes to the configuration, follow
the steps below:
1. Select [Start => Programs => IBM Dynamic IP Client] and then click
Dynamic IP Configuration.
2. Type the hostname, domain name, and primary domain name server (fully
qualified domain name) and then click OK. See Figure 103 on page 146.
3. Reboot your computer.
4. To check the return code of the NSUPDATE, open the Dynamic IP
Configuration and click on the Diagnostic tab. See Figure 104 on page
147.
5. To check the current configuration, type WINIPCFG at the command prompt.
When running on Windows 95, DDNS updates only occur automatically at the
correct DHCP renewal intervals. Manual renewal and release of the lease
using WINIPCFG.EXE does not coincide with an automatic DDNS updates.
When running on Windows 95, you must reboot the machine after changing
the configuration of the IBM Dynamic IP Client because the configuration
changes and updates will not occur automatically.
Tips for setting up
In 4.3.3, “OS/2 Warp 4 using presecured mode” on page 203, we discussed
that after importing a key file, the hostname, domain name, and primary
DDNS server name were automatically filled in for you in the configuration
notebook. However, using the IBM Dynamic IP client for Windows 95/NT, this
information is not filled in automatically (because there is no function to
import a key file).
148
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The information below describes our actions for a Windows 95 client. We
used one Windows 95 computer with IBM Dynamic IP Client installed and
used its registry information. We exported the registry file, modified the
configuration for each one of the other Windows 95 computers, and
distributed the appropriate registry file to them:
1. Open the registry.
2. Keys and values to modify are as follows:
HKEY_LOCAL_MACHINE/SOFTWARE/IBMDHCP/CLIENT/1.0/APPLICATION DATA
Set a DDNS server’s IP address or a hostname (fully qualified name) in
PrimaryDNS:
HKEY_LOCAL_MACHINE/SYSTEM/CURRENTCONTROLSET/SERVICES/VXD/MSTCP
1. Set domain name to Domain.
2. If EnableDNS is set to 0 (which means disabled), set the value to 1.
3. Specify a hostname of the client computer in Hostname.
3. Highlight MSTCP in the trees and select the Registry menu and Export
Registry File. This will create the <file_name>.REG file on the desktop.
Exported registry files are text files. You can edit the file using an editor for
other clients.
4. Export IBMDHCP in the same way.
5. On the target client, double-click on these registry files. This will register
the information in the file into the registry.
Figure 105. [Windows 95] Registry Editor message
6. Then, install the IBM Dynamic IP Client on the target machine. When the
configuration notebook comes up at the end of the installation, it already
contains all the information needed.
4.1.5.4 Windows 95 using proxy
The DHCP server updates both forward and reverse mappings of the client’s
hostname and IP address. The only configuration needed on the client
computer is to enable a DHCP client. To enable a DHCP client, refer to 2.5.2,
“Windows 95 and Windows 98” on page 72.
Chapter 4. Serving names
149
Windows 95 DHCP client sends its computer name (NetBIOS name) as
Option 12, and a DHCP server uses this name to update the DDNS server.
4.2 OS/2 Warp Server as Dynamic DNS server
In Chapter 2, “Up and Running with DHCP” on page 29, we discussed how to
set up and run a DHCP server. With a DHCP server in your network, IP
addresses and other configuration information are assigned automatically to
the clients. However, one problem persists:
• How can a domain name server learn about those dynamically assigned
IP addresses and hostnames so it can update its database accordingly?
One solution is to use a WINS server as discussed in 4.1.2, “Windows NT as
a DDNS using WINS” on page 136. However, if you have clients other than
Microsoft clients, they cannot register their names with WINS servers and
they cannot be reached by names. This problem can be solved by using the
Dynamic Domain Name System (DDNS) services.
This chapter shows you how to set up your OS/2 Warp Server as a Dynamic
DNS server. In our scenario, we describe OS/2 Warp Server for e-business,
which includes TCP/IP Version 4.21. TCP/IP 4.1 for OS/2 is similar. Before we
show you how to set up the server, we provide a brief introduction of DDNS in
the following section.
4.2.1 The Dynamic Domain Name System (DDNS)
The DNS servers only support queries on a statically configured database,
and the DNS administrators have to manually update this database whenever
there is a change in the network and client configurations. The Dynamic DNS
(DDNS) protocol defines an extension to this domain name system to update
the DNS database dynamically. Using DDNS protocol, both the address (A)
record, which maps a hostname to its IP address, and the pointer (PTR)
record, which maps a host’s IP address to its hostname, can be added and
deleted dynamically.
4.2.1.1 IBM Dynamic Domain Name System
The IBM Dynamic Domain Name System (DDNS) is based on, and is a
superset of, the Internet Software Consortium’s publicly available
implementation of the Berkley Internet Name Domain (BIND) Level 4.9.3.
DDNS supports both static and dynamic DNS domains. In dynamic domains,
only authorized clients can update their own data. In the IBM DDNS, RSA
public-key digital signature technology is used for client authentication.
150
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
With the IBM DHCP server, an administrator can configure host configuration
parameters only at a server and can automate the configuration of IP hosts.
DDNS provides dynamic hostname-to-IP address (and IP
address-to-hostname) mapping for Dynamic IP clients. Using DDNS, clients
automatically update their A record with their new address, obtained from the
DHCP server, and IBM DHCP servers can automatically update PTR records
for those clients.
4.2.1.2 DDNS mechanism
With DDNS, both address (A) records and the pointer (PTR) records can be
updated dynamically. The DDNS client program, NSUPDATE, is used to
update information in a DDNS server.
Dynamic updates are performed by any of the following:
• Network client: A host which has DHCP and DDNS client software and
can update its A and TXT records with the current IP address information.
• DHCP server: A host that updates PTR records with current hostname
information for the address it allocates, and on certain occasions, updates
A records for clients that either cannot or do not update the A records
themselves.
• DDNS system administrator: A user who can update data such as A and
PTR records in a dynamic domain.
The interaction between a DDNS client (including DHCP server) and a DDNS
server can be described as follows and is summarized in Figure 106 on page
152.
Chapter 4. Serving names
151
Figure 106. [Warp Server] DHCP – DDNS – Client interaction
• After a DHCP/DDNS client has been provided with all the information by a
DHCP server, it contacts a DDNS server.
• The DDNS client then sends an update request for the resource records
that are associated with the client’s hostname. The client also sends its
public key and signs all resource records with a digital signature. The key
and the signature, together, will allow the server to verify the authenticity
of the update, as well as any other software which supports the MD5
algorithm. For more information, see 8.2, “RSA public key authentication
system” on page 348.
• If the client’s updates have been successful, the server commits the
changes to its database, and the client becomes reachable by its
hostname.
• A name server normally supports inverse queries. This means the DDNS
server must be updated with the reverse information. A DHCP server
running on the OS/2 Warp server can update the reverse mappings for
clients. After the DDNS client has successfully registered its name with the
DDNS server, the DHCP client will send a lease renewal request message
to the DHCP server. The client will include the new hostname in the
message. The DHCP server is told to update the database with the client’s
new IP address and hostname.
152
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
4.2.1.3 Security
By having a DHCP and DDNS server in your network, hostnames can be
registered automatically and updated dynamically. However, this means that
without DDNS client authentication, an unauthorized host might impersonate
an unsuspecting host by remapping the address entry for the unsuspecting
host to that of its own and could intercept data (for example, logon
passwords) intended for the unsuspecting host.
RSA digital signature technology is implemented for DDNS in TCP/IP for
OS/2 to authenticate the owner of the DNS records and to secure the DNS
database updates.
RSA digital signature technology
RSA digital signature technology is based on the use of private-public key
pairs:
• The private key is used to generate a unique digital signature that can be
validated only by using the corresponding public key.
• The private key cannot be derived from either the digital signature or the
public key or the combination of the two.
When a DDNS client registers its hostname for the first time, it generates an
RSA key pair. A public key is sent to the DDNS server and registered for that
particular hostname.
The DDNS client retains the RSA key pair (with the private key encrypted) in
the client key file called DDNS.DAT (in the directory specified by the ETC
environment variable). When the DDNS client updates the resource records
on the DDNS server, the digital signature is generated by using the private
key and is sent to the server with the update data.
When the DDNS server receives the update request, it uses the DDNS
client’s public key (in the KEY resource record) to:
• Authenticate the owner of the update request to verify that the update
request was signed with the corresponding private key.
• Verify that the data was not changed since it was signed. (In other words,
to verify that no one intercepted and changed the data on the way to the
name server.)
In this way, only the owners of the original records, who posses the correct
private key necessary to generate the correct digital signature, can update
resource records that are protected by an existing KEY resource record.
Chapter 4. Serving names
153
In addition, a DDNS administrator can create and use a zone key for each
dynamic domain. The zone key is the administrator’s RSA key pair for a
particular domain. This key enables the administrator to create, modify, or
delete any host’s record in the domain, regardless of who created the
records. The private key is used to generate a signature when the
administrator updates the requests, and the server examines the signature to
verify that update requests are from the administrator. The public key is
registered in the domain data file as a KEY resource record for the domain.
The private and public keys are stored in the DDNS.DAT file as the
administrator’s key file in the directory specified by the ETC environment
variable.
For more information, see 8.2, “RSA public key authentication system” on
page 348.
Security planning
As discussed in the previous section, DDNS servers use RSA digital
signatures to authenticate DDNS updates requests. DDNS supports two
modes of securing updates for a dynamic domain. Both modes protect the
records in the database from unauthorized updates.
154
Dynamic Secured Mode
This is a default mode. In the dynamic secured
domain, DDNS clients can generate their own
RSA key pair and dynamically register a KEY
resource record containing the public key
when they create their own name record for
the first time. This means that servers allow
any hosts with DDNS protocol to register their
hostnames, however, once registered, only
registered hosts (and administrators) are
allowed to update their own entry.
Dynamic Presecured Mode
This is an alternate mode to the secured
mode. Figure 107 on page 155 summarizes
dynamic presecured mode. In the presecured
dynamic domain, DDNS clients must be
pre-authorized by a DDNS administrator
before they can create their name record. The
DDNS administrator must preregister hosts
and generate RSA keys for each client. This
means that in presecured mode, the KEY
resource record must be already defined in the
domain before an update is accepted. The
DDNS administrator must also distribute the
correct corresponding key information to each
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
host before they create the resource records.
See 8.4, “Presecured domain” on page 353,
for more information.
Figure 107. [Warp Server] Dynamic Presecured Domain
For clients that do not have DDNS Client function, you can use a function
called ProxyArec.
ProxyArec
For a client that can not update its A record by
itself, a DHCP administrator can configure the
DHCP server to update both the A and PTR
records on behalf of the clients, as shown in
Figure 108 on page 156. This function is
called a ProxyArec. When using ProxyArec, a
DHCP server uses a hostname sent by a client
through Option 12 to update a DDNS server.
For more information, refer to 8.5, “ProxyArec
consideration” on page 356.
Chapter 4. Serving names
155
Figure 108. [Warp Server] using ProxyArec
4.2.2 The IBM Dynamic DNS (relative to other implementations)
IBM implemented Dynamic DNS in 1995. This implementation was done by
IBM research based on an Internet draft being discussed in the IETF at that
time. The protocol was soon going to become standard.
4.2.2.1 Basic design
The basic design involves the following items:
• Allowing updates to the DNS by remote client machines.
• Having these updates be authenticated using RSA digital signatures, as
illustrated in “RSA digital signature technology” on page 153.
• Having the concept of expiration of resource records, so when an update
was done in coordination with DHCP, the entries expired along with the IP
address lease, thus making the system self administering.
There were also two RSA public-private key management mechanisms in the
code:
• The secured mode, where the client could create/register its public key in
the DNS with a hostname, and after that point, be the only authorized
entity to update that information. You can find detailed information on this
topic on page 154.
156
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• The presecured mode, where the systems administrator was the only
authorized entity to create/register a public key with a hostname. After that
time, the client machine that had the corresponding private key was the
authorized entity to update that entry. There is more information on this
topic on page 154.
The system was designed to integrate DHCP and DDNS, and more
importantly, allow the client to manage its own A record in the DNS.
It is important to note that the IETF only standardizes protocol on the wire.
The key management policies, the integration of DHCP and DDNS are all IBM
design, which differentiate an integrated product from a pure technology.
4.2.2.2 Current state of the standards
The IETF did not standardize the draft that IBM research worked from. That
standard evolved into the current RFC 2137 (Secure Dynamic Updates). Its
resource records in the DNS are almost identical to the IBM implementation.
However, the protocol on the wire is different. It also deals with the issue of
integrating dynamic updates into secure DNS resolving (RFC 2065) and
describes two modes of operation, mode A and mode B. In mode A, the
client-signed RRs are stored as part of the DNS data, whereas in mode B, the
server signs all the records. The IBM implementation of dynamic updates is
actually very close to mode A (although the secure resolving described in
RFC2065 is not supported yet).
The IETF also has an unsecure updates standard RFC 2136. (It is currently in
BIND 8.1.)
There is also another Internet draft using Transaction Signatures (T-Sig),
which uses shared secrets rather than public key cryptography to secure
communication between the nameserver and a client. It seeks to be a simpler
alternative to both RFC2065 for secure resolving and RFC2137 for dynamic
updates in those situations where the client and the server can establish a
trust relationship. This is the mechanism Microsoft will probably use in
Windows NT 5.0.
4.2.2.3 Differences in standards
The following differences apply to IBM’s implementation of secured Dynamic
DNS versus RFC 2137, RFC2136, and T-Sig.
Chapter 4. Serving names
157
• IBM DDNS and RFC 2137
In normal cases, the logical evolution of IBM implementation would be to
move towards the RFC 2137 mechanism. This would not lead IBM to any
new function, but IBM would have the comfort of stating compliance with
RFC 2137. This would become important if other vendors were developing
clients or servers with this capability. No one has done so until now; so,
doing this does not presently give IBM any practical advantage over
maintaining the current method.
• IBM DDNS and RFC2136
RFC 2136 was not designed for any authentication or expiration of
records. It was also not designed for clients updating the DNS records. It
does not have the following capabilities, which were essential for our
integrated product:
- No real security mechanism. It accepts a list of IP addresses where the
updates to any record in the system can originate (per zone). Thus, not
only is this mechanism unsecure (anyone who can update can update
anyone's record), it is not scalable. One cannot realistically list every
possible IP address all the clients may be sending. Some users have
tried to use this in a proxy-update mode, where the DHCP server
updates both the entries, and the DHCP server's IP address is the only
one allowed to update. This mechanism still has security flaws, since
any machine can send any other machine’s hostname as the requested
name, and the DHCP server has no mechanism of verifying that
information.
- No mechanism of expiration. Thus, if a DHCP client got an address and
then moved away from the network, the DHCP server's IP address
lease would expire after a period of time and the address would get
reused, but the DNS entry would have to be manually removed.
- No mechanism to detect out-of-order update requests or prevent
replay-attacks. (No timestamp associated with the update request.)
• IBM DDNS and T-Sig
T-Sig accomplishes the goal of providing an authentication mechanism for
updates from clients. But, it still does not provide an expiration
mechanism. This means that a client created its A-RR and moved away
from the network for some period of time. The PTR record created by the
DHCP server could be deleted by it, but the A-RR (the hostname to IP
address mapping) would remain in the DNS. So, if someone else
connected to the network and got assigned the same IP address by the
DHCP server (which is legal since the lease expired), that machine also
received a hostname that it did not have access to.
158
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
T-Sig also has scalability issues, since the shared secret keys are
theoretically unique between every client-name server pair, and must be
predefined in each.
T-Sig does have a timestamp associated with its update requests, but
since these timestamps are not stored with the data updated, it is
questionable how out-of-order update requests or replay attacks could be
effectively detected and/or prevented.
4.2.2.4 Conclusion
Moving to the BIND 8.1.1 or the T-Sig capability introduces flaws in the
update system, and is not recommended from a system design perspective.
4.2.3 Scenario
In this section, we give an example of a simple network environment that
includes only one subnet. In this network, we have a DHCP server on OS/2
Warp Server for e-business (See 2.2, “OS/2 Warp Server for e-business as a
DHCP server” on page 30 for details on DHCP server setup) running and
another OS/2 Warp Server with WorkSpace On-Demand manager running.
We also have a few DHCP/DDNS client computers and WSOD clients. When
we started building our network, we set up a DHCP server, and, initially, did
not have a DNS/DDNS server. In this section we will integrate DDNS into our
network.
In this scenario, we configure OS/2 Warp Server as a Dynamic DNS server.
We configure a DDNS server using one of two methods:
1. Dynamic secured mode
2. Dynamic presecured mode
Later in this chapter, we also configure our DHCP server to enable ProxyArec
and have Windows 95 and WSOD clients configured through this function.
In our example, our domain name is armonk.cooking.net, and the
DHCP/DDNS server's hostname is merlot. Table 7 shows our network
configuration:
Table 7. OS/2 Warp Server configuration information
Configuration settings
Value
Network ID
192.168.6.0
Subnetmask
255.255.255.0
Domain name
armonk.cooking.net
Chapter 4. Serving names
159
Configuration settings
Value
DHCP/DDNS Server IP address
192.168.6.10
DHCP/DNS/WINS server hostname
merlot
4.2.4 Tasks
The following list describes the tasks required to complete the scenario,
followed by detailed, step-by-step instructions.
1. Plan the primary domain.
2. Define your OS/2 Warp Server as a primary DNS server for your DDNS
domain.
3. Create a primary domain for forward mappings.
4. If using presecured mode, generate keys for the clients.
5. If there are hosts using a static IP address, add static records for them.
6. Configure a DHCP server to update PTR records.
7. Start the DDNS server
8. Restart the DHCP server.
9. Verify that the DDNS server is operating correctly.
4.2.5 Planning
Before you start configuring your Dynamic DNS server, you need to make a
plan for the design of your network that includes domain and security issues.
4.2.5.1 Designing your domain and domain name server
The first step is to design your domain. Before you establish DNS servers in
your network, you need to determine the domain name for your domain.
Furthermore, you need to determine which server is going to be the Dynamic
DNS server for your domain.
In our scenario, armonk.cooking.net is the primary domain name, and merlot
is the name of the name server. All hosts on the 192.168.6.0 network are
included in the domain armonk.cooking.net.
Then, you need to decide which hosts are going to be the DHCP/DDNS
clients, and which are not. A DDNS name server must have a static IP
address and, therefore, cannot be a DHCP client. If you have any other hosts
that have static IP configuration, you have to decide on their names and
configure the DNS database manually.
160
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
4.2.5.2 Security planning
The next step is to decide on security issues. Dynamic DNS in TCP/IP
Version 4.1 for OS/2 supports two modes of security:
• Dynamic Secured Mode
• Dynamic Presecured Mode
To decide which mode to implement, refer to 4.2.1.3, “Security” on page 153.
Also see Chapter 8, “Security of DHCP and Dynamic DNS” on page 347. If
you have clients that do not support DDNS, you need to enable the ProxyArec
function at the DHCP server. For more information about ProxyArec, refer to
8.5, “ProxyArec consideration” on page 356.
4.2.6 Setting up your DDNS server
This section describes how to set up a DDNS server on your Warp Server.
See 4.2.6.2, “Dynamic secured mode” on page 162 for details on how to set
up a DDNS server with dynamic secured mode, and 4.2.6.3, “Dynamic
presecured mode” on page 168 shows you how to set up with dynamic
presecured mode. All values we used for configuration in the following steps
are the values we defined in our scenario.
4.2.6.1 Before setting up
Before you start configuring, make sure that the TCP/IP is configured
correctly. The Dynamic DNS server must have a static IP address, and,
therefore, cannot be a DHCP client.
In addition, before using the DDNS Server Administrator GUI to configure
DDNS, make sure that you have created a TCP/IP Administrator password.
This was described in 2.2.2, “DHCP server preparation” on page 31.
The Dynamic DNS server itself can also be a DNS client. Follow the steps
below to add the DNS server information:
1. Open the TCP/IP Shadows folder and the TCP/IP Configuration folder,
then double-click on TCP/IP Configuration (Local) or type TCPCFG2 at an
OS/2 command prompt. The TCP/IP configuration settings notebook is
displayed.
2. Click on the Host Names tab. Type hostname and local domain name.
3. Click Add. Type an IP address for the computer (DNS/DDNS server) and
click on OK.
4. Close and save the configuration.
Chapter 4. Serving names
161
Important
The next step is very important. You need to create an entry in the HOSTS
file. DDNS Administration GUI tool uses information in the HOSTS file, and
this is required to automatically create a primary zone for reverse
mappings.
1. Open the HOSTS file using an ASCII editor. The HOSTS file is stored in
the directory specified by the ETC environment variable. (If you are
uncertain, type SET ETC to check which directory the ETC environment
variable points to.)
2. Add your name server hostname using the fully-qualified domain name.
Figure 109 shows you an example of the HOSTS file after adding a new
entry.
127.0.0.1
192.168.6.10
localhost
merlot.armonk.cooking.net
Figure 109. [Warp Server] HOSTS file
4.2.6.2 Dynamic secured mode
Follow the steps below to set up the DDNS server using dynamic secured
mode:
1. You will use the DDNS Server Administrator program to configure the
DDNS server. To do so, open the TCP/IP Shadows folder, then open
DDNS Server Services and double-click on the DDNS Server
Administrator icon.
2. DDNS Server Administrator window is displayed. Enter the TCP/IP
administrator password and click OK.
Two settings notebooks are displayed.
3. In the Domain Name Server settings notebook, click the Server tab. Verify
the name and IP address of the name server. (The GUI retrieves the
hostname and IP address from the HOSTS file.)
162
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 110. [Warp Server] Domain Name Server notebook
If desired, specify a different cache file name on the Files tab. Click OK to
close the settings notebook.
4. At the DDNS Server Administrator window, define the DDNS server as the
primary server for your domain as follows:
• Click on the Add Primary Domain button to open the Primary Domain
notebook.
• At the Domain Configuration tab, type your complete domain name.
Chapter 4. Serving names
163
Figure 111. [Warp Server] Primary Domain Configuration
• At the Domain Configuration tab, note that the domain type is Dynamic
by default, as shown in Figure 111. Click on Dynamic Options and
note that Client automatically creates its own hostname (which means
Dynamic Secured mode) is checked, as shown in Figure 112. Click OK
to go back to the Domain Configuration tab.
Figure 112. [Warp Server] Setting secured mode
164
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• If you do not want to create a reverse mappings file (by default), click
on the Domain Options tab and uncheck Automatically Create
Reverse Mappings for Statically Defined Hosts.
• If you want to change the name of the file for your domain, click on the
Files tab and type the path and the file name of the domain data file.
• After you finish the configuration, click OK to close the notebook. When
you close the Primary Domain Notebook, a zone key is automatically
created.
5. When you create a primary domain, an A record, PTR record, and an alias
ns-updates for the name server is created automatically, as shown in
Figure 113 on page 165. If you do not want to use this alias, delete the
alias or use the Alias Notebook to add a different alias.
Figure 113. [Warp Server] DDNS Server Administrator notebook
Chapter 4. Serving names
165
ns-updates alias
A DDNS client must know the fully qualified domain name or IP address of
the primary DDNS server. DDNS clients using the DDNS client
configuration program for a DHCP client (DDNSCFG.EXE) defaults the
name of the Primary DDNS server to ns-updates.<domain_name> (In our
scenario, ns-updates.armonk.cooking.net.) An end user can change this
ns-updates default at any time by using the DDNS client configuration
program. If you follow the steps to modify HOSTS file, as described in
4.2.6.1, “Before setting up” on page 161, this alias ns-updates is created
automatically.
6. If you want to add any static hosts, follow the steps below:
• On the DDNS Server Administrator notebook, highlight the domain in
which you are going to add a new host, and click Add Host on the tool
bar.
• Type the hostname and IP address of the computer and click Add.
Click OK when finished.
Figure 114. [Warp Server] Adding Hosts
7. Click Save from the File menu to save the configuration.
8. You need to configure the DHCP server so that DHCP clients can get the
configuration information for DDNS/DNS server and domain name. Add
the following two options:
• Option 6 Domain Name Server: IP address of the DNS/DDNS server
• Option 15 Domain name: Domain name
166
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Refer to 2.2, “OS/2 Warp Server for e-business as a DHCP server” on
page 30 to see how to add DHCP options.
9. If you want to update PTR records for DHCP/DDNS clients automatically,
configure your DHCP server, as described in the steps below:
• Open DHCP Server Service in the TCP/IP Shadows folder and
double-click on DHCP Server Configuration.
• Enter your password and click OK. The DHCP Server Configuration
window comes up.
• In the Current Configuration lists, double-click on your subnet to open the
subnet notebook.
• Click on the Miscellaneous tab, and type the IP address of the DDNS
server in the DDNS Server for PTR field.
Figure 115. [Warp Server] Specifying DDNS server for PTR record update
• Click OK to close the notebook.
Chapter 4. Serving names
167
• Repeat steps above for all the subnets you want to update PTR records
dynamically.
• In the Current Configuration lists, double-click on DHCP Server to open
the DHCP Server Parameters notebook.
• Click on the DDNS PTR Records tab, check Automatically update or
delete PTR records and click OK.
Figure 116. [Warp Server] Enabling DHCP Server to update PTR records
• Select [File —Save] to save the configuration in the DHCP Server
Configuration window.
10. Double-click DDNS Server in the DDNS Server Service folder to start the
DDNS server.
11. Restart the DHCP server.
4.2.6.3 Dynamic presecured mode
In this section, we show you how to setup a dynamic presecured domain
using the DDNS Server Administrator program.
168
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
1. Open the TCP/IP Shadows folder, then open DDNS Server Services, and
double-click on the DDNS Server Administrator icon.
2. The DDNS Server Administrator window comes up. Type the password
and click OK.
3. Two notebooks come up. On the Domain Name Server notebook (Figure
117), click on the Server tab. Verify the name and IP address of the name
server. (GUI uses the hostname and IP address from the HOSTS file.)
Figure 117. [Warp Server] Domain Name Server notebook
If desired, on the Files tab, specify a different cache file name. Click OK to
close the notebook.
4. Using the DDNS Server Administrator window, define the DDNS server as
a primary server for your domain as follows:
• Click Add Primary Domain on the tool bar to open the Primary
Domain notebook.
• On the Domain Configuration tab, type your complete domain name.
Chapter 4. Serving names
169
Figure 118. [Warp Server] Primary Domain Configuration notebook
• On the Domain Configuration tab, note that the Domain Type is set to
Dynamic, as shown in Figure 118. To set the mode to presecured,
press Dynamic Options to go to the Dynamic Options Notebook and
then check Only System Administrator Can Create Client’s Host
Name (dynamic presecured mode), as shown in Figure 119.
Figure 119. [Warp Server] Setting Dynamic presecured mode
170
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• If you do not want to create a reverse mapping (by default), click the
Domain Options tab and uncheck Automatically Create Reverse
Mappings for Statically Defined Hosts.
• If you want to change the name of the file for your domain, click on
Files tab and type the path and file name of the domain data file.
• After you finished your configuration, click OK to close the notebook.
When you close the notebook, the zone key for the domain is created
automatically.
5. When you create a primary domain, an A record, PTR record, and an alias
ns-updates for the name server is created automatically, as shown in
Figure 120. If you do not want to use this alias, delete the alias or use the
Alias Notebook to add a different alias.
Figure 120. [Warp Server] DDNS Server Administrator notebook
Chapter 4. Serving names
171
ns-updates Alias
A DDNS client must know the fully qualified domain name or IP address of
the primary DDNS server. DDNS clients using the DDNS client
configuration program for a DHCP client (DDNSCFG.EXE) defaults the
name of the primary DDNS server to ns-updates.<domain_name>. (In our
scenario, ns-updates.armonk.cooking.net.) An end user can change this
ns-updates default at any time by using the DDNS client configuration
program. If you follow the steps to modify HOSTS file, as described in
4.2.6.1, “Before setting up” on page 161, this alias ns-updates is created
automatically.
172
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
6. If you want to add any static hosts, follow the steps below:
• In the DDNS Server Administrator window, highlight the domain in
which you want to add a new host and click Add Host on the tool bar.
• Type the hostname and the IP address of the computer and click Add.
Click OK to finish the window.
Figure 121. [Warp Server] Adding Hosts
7. Click Save from the File pull-down menu to save all configuration
information.
8. Double-click on the DDNS Server object in the DDNS Server Service
folder to start the DDNS server.
9. The last step is to register DDNS clients and generate host keys for them
(the DDNS server must be started to generate keys):
• Start the DDNS Server Administrator program.
• Double-click the domain for forward mappings in the list to open the
Primary Domain notebook.
• On the Domain Configuration tab, click on the Dynamic Options
button.
• Click Generate Keys for Hosts on the Dynamic Options Notebook.
• On the Generate Keys Window, type hostnames and click Add for each
client. If you want to change the directory where keys are stored,
specify the path. Click OK to return to the Dynamic Options Notebook.
Chapter 4. Serving names
173
Figure 122. [Warp Server] Generate Keys for DDNS clients
10. Click OK to close the Dynamic Options settings notebook.
11. Click OK to close the Primary Domain settings notebook. When you close
the Primary Domain settings notebook, the host keys that you registered
are created automatically under the directory you specified.
12. You need to configure the DHCP server so that DHCP clients can get the
configuration information for the DDNS/DNS server and domain name.
Add the following two options:
• Option 6 Domain Name Server: IP address of the DNS/DDNS server
• Option 15 Domain name: Domain name
Refer to 2.2, “OS/2 Warp Server for e-business as a DHCP server” on
page 30 to see how to add DHCP options.
13. If you want to update PTR records for DHCP/DDNS clients automatically,
configure a DHCP server as described in the steps below:
• Open DHCP Server Service in the TCP/IP Shadows folder and
double-click on DHCP Server Configuration.
• Enter your password and click OK. The DHCP Server Configuration
window comes up.
• In the Current Configuration lists, double-click on the subnet <subnet
name> to open the subnet notebook.
• Click on the Miscellaneous tab and type an IP address of the DDNS
server in the DDNS server for the PTR Record update field.
174
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 123. [Warp Server] Specifying DDNS server for PTR record update
• Click OK to close the settings notebook.
• Repeat the above steps for all the subnets you want to update PTR
records dynamically.
• In the Current Configuration lists, double-click on DHCP Server to open
the DHCP Server Parameters settings notebook.
• Click on the DDNS PTR Records tab, check Automatically update or
delete PTR records, and click OK.
Chapter 4. Serving names
175
Figure 124. [Warp Server] Enabling DHCP Server to update PTR records
• Select [File => Save] to save the configuration in the DHCP Server
Configuration window.
14. Restart the DDNS server.
15. Restart the DHCP server.
4.2.7 ProxyArec
If you have any DHCP clients that do not also have the DDNS client function,
you can configure your DHCP server to update the A (address) records on
behalf of the DHCP clients.
To enable ProxyArec on your DHCP server, see 8.5, “ProxyArec
consideration” on page 356.
176
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
4.2.8 Verification
The last step is to make sure that your DDNS server is working properly. To
verify the configuration, you can do the following:
• Verify that the name server can resolve names.
• Check the NSUPDATE log file.
4.2.8.1 Verifying the DDNS server configuration
Use the HOST and NSLOOKUP commands to check whether the DDNS server can
resolve names.
The HOST command contacts a DNS server to translate a specified hostname
to its IP address or specified IP address to its hostname. For syntax
information on the HOST command, refer to 10.2.6, “The HOST command” on
page 388.
The command NSLOOKUP queries a DNS server in either interactive or
non-interactive mode. The NSLOOKUP command syntax information is provided
in 10.2.7, “The NSLOOKUP command” on page 389.
4.2.8.2 Checking the NSUPDATE log
The NSUPDATE agent, when used for a DHCP client, or server, logs all
significant events to the NSUPDATE.LOG file. This file is stored in the
directory specified by the ETC environment variable.
4.2.9 DDNS files
This section describes the configuration files used by the DDNS server. The
configuration files used by the DDNS server are:
• Configuration file
• Boot file
• Cache file
• Domain data files
These files are created and updated automatically when you configure your
domain using the DDNS Server Administrator program. It is possible to create
these files manually, but we recommend you use the DDNS Server
Administrator program, because this program checks the configuration to see
if there are any mistakes in the configuration you defined. It also
automatically updates the serial number so that the secondary name server
can notice that there has been a change in the database. In addition, by using
Chapter 4. Serving names
177
a GUI, you do not need to execute the DDNSZONE command to generate keys,
because the program automatically does it for you.
4.2.9.1 Configuration file DNSEXT.CFG
Many of the new enhancements IBM has added to the standard BIND DNS
code require configuration. These new features are configured in a separate
file, \MPTN\ETC\NAMEDB\DNSEXT.CFG, so that there is no confusion with
standard name server configuration files.
The DDNS server GUI creates the DNSEXT.CFG file. It has a few entries that
it places in there for its own information. However, this file is completely
optional, since all parameters have default values.
Figure 125 on page 179 shows an example of the DNSEXT.CFG file.
178
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
; *************************** IBM DDNS Server Administrator ******************
; This file was written by the IBM DDNS Server Administrator on 28-Jun-98
; *************************** IBM DDNS Server Administrator ******************
austin.cooking.net (
notify=yes
notify.add=192.168.6.2
;notify.add=192.168.7.2
(This line is commented)
notify.delayTime=60
notify.retryTime=30
notify.retryNumber=3
timeSync=yes
timeSync.toSecondaries=yes
safeWrite=yes
sigDel=no
ttlSet=no
deferUpdCnt=100
incrTime=300
keyToSec=yes
sepDynStatic=yes
reverseMapping=yes
;Used by the DDNS GUI only)
)
6.168.192.in-addr.arpa (
notify=yes
notify.delayTime=60
notify.retryTime=30
notify.retryNumber=3
timeSync=yes
timeSync.toSecondaries=yes
safeWrite=yes
sigDel=no
ttlSet=no
deferUpdCnt=100
incrTime=300
keyToSec=yes
sepDynStatic=yes
reverseMapping=yes
)
DDNSAdministratorClient (
gui.warn=yes
gui.write=yes
gui.num=100
gui.lease=3600
gui.pad=3110400
gui.reinit=1
gui.sepdata=3
)
Figure 125. [Warp Server] DNSEXT.CFG
Some of the configuration parameters used in the DNSEXT.CFG file are
described in Table 8 on page 180.
Chapter 4. Serving names
179
Table 8. DNSEXT.CFG Configuration Parameters
180
Configuration
parameters
Parameter
Values
Description
Notify
parameters
notify
yes
no
All notify parameters apply
to both secondary and
primary zones. This
parameter must be set to
yes in order for the other
notify parameters listed
below to apply.The default
is no.
notify.add
IP
address
Optional parameter that
can be used to add a
particular DDNS server’s
IP address to the list of
servers to notify.
notify.remove
IP
address
Optional parameter that
can be used to remove a
particular DDNS server’s
IP address from the list of
servers to notify.
notify.delayTime
number
of secs
This is the amount of time
between notifications sent
to consecutive servers on
the notify list. The actual
delay time is a random
time between the specified
value and twice that value.
The default value is 30.
notify.retryTime
number
of secs
This is the amount of time
between notifications sent
to a particular server.
(Multiple notifies will be
sent to a server if that
server does not respond.)
The default value is 60.
notify.retryNumber
number
This is the maximum
number of times that a
notification will be retried
to a particular server. The
default value is 3.
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Configuration
parameters
Parameter
Values
Description
Client/server
time
synchronization
parameters
timeSync
yes
no
This parameter defines
whether the time
synchronization feature is
enabled or not, and it
applies only to primary
dynamic zones. The
default value is yes.
timeSync.toSecondaries
yes
no
This parameter defines
whether the time
synchronization being kept
should be forwarded to the
secondary servers or not.
This parameter will be
ignored if the timeSync is
set to no. The default value
is yes.
sigDEL
yes
no
This parameter defines
whether the signature
deletion function, which
deletes signature records
for expired host names,
should be enabled (yes) or
not (no), and applies only
to the primary dynamic
zones. The default value is
no.
sigDEL.time
number,
(0 -23)
military
time
This parameter defines the
military time at which any
signature records still
remaining for the zone will
be deleted. This
parameter is ignored if
sigDEL is set to no.
safeWrite
yes
no
This parameter defines
whether the safe write
feature, which keeps a
backup of the zone data,
should be enabled (yes) or
not (no). The default value
is yes.
SIG RR
deletion
parameters
Safe write to
disk for zone
data
parameter
Chapter 4. Serving names
181
Configuration
parameters
Parameter
Values
Description
Override TTL
for dynamic
data
parameters
ttlSet
yes
no
This parameter defines
whether the TTL override
function, which allows
administrators to lower the
TTL value for dynamic
clients, should be enabled
(yes) or not (no). This
parameter applies only to
primary dynamic zones.
The default value is no.
ttlSet.value
number
The number is the actual
TTL value to use in the
override of dynamic client
TTLs. This parameter will
be ignored if ttlSet is set
to no. It is required if
ttlSet is set to yes. The
valid numbers for a TTL
are the positive values of a
signed 32-bit number.
sepDynStatic
yes
no
This parameter defines
whether the
static/dynamic data
separation feature, where
static and dynamic data for
a dynamic zone are kept
separately, is enabled
(yes) or not (no). This
parameter applies only to
primary dynamic zones.
Separate
static /
dynamic zone
data
parameters
182
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Configuration
parameters
Parameter
Values
Description
Configuration
parameters for
Interoperability with
earlier
versions of
DDNS
keyToSec
yes
no
If keyToSec is set to no, it
corresponds to
nokeytosec on the
primary line of the boot file
in earlier versions of
DDNS.
incrTime
number
of
seconds
The default value is 300
seconds (five minutes).
This value corresponds to
the optional sixth
parameter of the SOA RR
for primary dynamic zones
from earlier releases of
DDNS. It specifies the
maximum amount of time
to wait after an update
before incrementing the
zone serial number.
deferUpdCnt
number
The default value is 100.
This value corresponds to
the optional seventh
parameter of the SOA RR
for primary dynamic zones
from earlier releases of
DDNS. It specifies the
maximum number of
updates that can occur
before the zone serial
number is incremented.
reverseMapping
yes
no
This attribute is placed
inside normal domain
entries.
Parameters
used by the
DDNS Server
GUI.
DDNSAdministratorClient
This value is set up in the
form of a domain that is
used by the GUI. This
domain and all its entries
are created and used
exclusively by and for the
GUI.
Chapter 4. Serving names
183
4.2.9.2 Boot file
This file is read by the name server when it is started. The default name of
this file is NAMED.BT or NAMED.BOOT and stored in the
\MPTN\ETC\NAMEDB directory. This file includes information such as the
domain names for which this name server is authoritative and the location of
the domain data.
Figure 126 shows an example of the NAMED.BT file.
;************************** IBM DDNS Server Administrator ************************
; This file was written by the IBM DDNS Server Administrator on 28-May-98
;************************** IBM DDNS Server Administrator ************************
primary 6.168.192.in-addr.arpa C:\\MPTN\\ETC\\namedb\\dnsf0000.rev dynamic secured
primary armonk.cooking.net C:\\MPTN\\ETC\\namedb\\dnsf0000.dom dynamic secured
cache . C:\\MPTN\\ETC\\namedb\\named.ca
Figure 126. [Warp Server] NAMED.BT file
4.2.9.3 Cache file
The cache (root server) file contains the names and IP addresses of the root
servers. The path and name of this file are specified in the cache entry in the
boot file.
4.2.9.4 Domain data file
Domain data files contain information about a domain, such as the IP
addresses and names of the hosts in the domain.
The forward domain data file contains entries that map hostnames to IP
addresses for hosts in the domain. The file also contains other data for the
hosts, such as mail (MS) or text (TXT) resource records. For dynamic
domains with separated static and dynamic data, the forward domain data
file, which contains dynamic data, must contain an $INCLUDE statement to
include the file that contains the static data. For example:
$INCLUDE c:\\mptn\\etc\\namedb\\static.data static.aromonk.cooking.net
The reverse domain data file contains entries that map IP addresses to
hostnames.
The domain data files can be named as desired, and the chosen name must
be put in the NAMED.BT file or specified in the $INCLUDE directive.
Figure 127 and Figure 128 on page 185 show the sample forward domain
data files for a dynamic secured domain with separated files for dynamic data
(DNSF0000.DOM) and static data (DNSF0000.STA).
184
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
;************************* IBM DDNS Server Administrator **************************
;This file was written by the IBM DDNS Server Administrator on 28-May-98
;************************* IBM DDNS Server Administrator **************************
$ORIGIN cooking.net.
armonk
IN
KEY
0x0080
0
1 AQOhHKv3ZhrCc7Hgsq2FglBW0t4qKWv+Gjmf
VQ8sd21kCVc6NcNv1QNWex21UedaE7AmZSZhKV5FSM9pPqqx8Rkd ;Cl=3
IN
NS
merlot.armonk.cooking.net.
;Cl=3
IN
SOA
merlot.armonk.cooking.net. administrator.armonk.coo
king.net. (
12 10800 3600 604800 86400 )
;Cl=3
$ORIGIN armonk.cooking.net.
pizza
3600
IN
KEY
0x0000
0
1 AQPTXIKlHuTRxQzSS7qUAdPndTt17VCr8b1m
Yo/ZQ72/Q+nkpwoia5NNLmB66G0FVJh62IOYAP8QQ7oZdq0lNMWv ;Cr=auth
4660
IN
A
192.168.6.12
;Cr=auth
4660
IN
SIG
A 1 4 4660 896410050 896323650 0xc534 pizza.armonk.
cooking.net RgZRe5FQ6Fv3vygfT0Xrtq1lZqTSi3d8fw/qjDLqHBLWMe162YRo8xqmuALY8IOHX/jKCpF
aEBFq9wgt5SfCxA== 896460438
;Cr=auth
4660
IN
SIG
KEY 1 4 4660 899520450 896323650 0xc534 pizza.armon
k.cooking.net Kd+Y7p92VvGy5XD5TYHmEwvc5BknB6xd/9G1AAIqFk2NpXewSJSeW4/anLgz2xT6tT3m5
dEsjjgbB6K8/VNbrQ== 899570838
;Cr=auth
burger 3600
IN
KEY
0x0000
0
1 AQOyn4i0WzIt6io3CQOaT7kbqLuFAMBLAhY7
BN7nNnQqy5DCmXD2lac386FgwGds4VwzI6RBYZmDUFyMZXTJSJSH ;Cr=auth
4660
IN
A
192.168.6.13
;Cr=auth
4660
IN
SIG
A 1 4 4660 896502341 896459141 0x9448 burger.armonk
.cooking.net ebko2wARtFhFxPMEm/Yb69+BPcmx6vQ2hqbfwrlmtmDAspnraXVGVX+V17PRPQoxWlYypk
0UOMhwL5Pj9TX05A== 896495156
;Cr=auth
4660
IN
SIG
KEY 1 4 4660 899612741 896459141 0x9448 burger.armo
nk.cooking.net PaJ4+IBi+vYBclxyBsC6WQRNOsxrekECerFCjZFyjEVCpBArDR05nlNvdTB0z2O/2oed
+8bMc66cmBR1w9Zo4w== 899605556
;Cr=auth
$INCLUDE C:\\MPTN\\ETC\\namedb\\dnsf0000.sta armonk.cooking.net.
Figure 127. [Warp Server] DNSF0000.DOM file
Note: Lines that are longer than 80 characters are wrapped to the next line.
;************************* IBM DDNS Server Administrator ***************
;This file was written by the IBM DDNS Server Administrator on 28-May-98
;************************* IBM DDNS Server Administrator ***************
$ORIGIN armonk.cooking.net.
ns-updates IN
CNAME merlot.armonk.cooking.net.
merlot IN
A
192.168.6.10
Figure 128. [Warp Server] DNSF0000.STA file
4.2.10 Providing mail services
The following discussion pertains to installing mail services on an OS/2 Warp
Server using Microsoft and IBM mail clients. The DHCP options and DNS MX
record discussion are applicable to any server platform. Only GUI use is
platform specific.
Now, we have automatic IP addresses, and our Dynamic DNS server is
keeping track of names for us. One of the first tools to set up now is the mail
system. We are going to set up mail services on our OS2 Warp Server. There
Chapter 4. Serving names
185
is a very robust POP3 mail server available called OS2POPS. This was
written by IBM Employees and is available for free distribution and use. We
have included Version 2.02 on the CD-ROM, but for the latest version, see:
http://www.raleigh.ibm.com/misc/os2pops
The SMTP mail server (sendmail) is shipped with OS/2 already; so, you can
have a mail system for no extra money. If you have decided to use Domino,
you will probably want to use the SMTP (Simple Mail Transport Protocol) and
POP3 (Post Office Protocol) mail servers included in that product (refer to
your Domino documentation for this). The DHCP and DNS records we will talk
about will be applicable to your Domino installation, as well as you mail
system installation.
In this section, you will learn how to configure a mail server so that it works
with your existing DNS system. You will learn how to install, configure, and
support the mail system. We will use the DHCP GUI to provide our mail
server address to our clients. We will also use the nslookup tool to examine
our name server records. Currently, if you run nslookup on your name server,
you will not see any MX (mail exchanger) records. We will add the MX record
after we setup the mail server. As mentioned in 1.6.5, “Resource record
types” on page 23, an MX record indicates a host that will deliver the mail to
the addressee, or it will forward the mail to another host that can do the same
(that is, deliver or forward the mail). Forwarding is accomplished via SMTP. To
summarize, the necessary steps are:
• Install, configure, and run a POP3 mailer.
• Install, configure, and run an SMTP mailer.
• Configure the DDNS server for mail processing or forwarding (MX
records).
• Configure DHCP server to provide mailer addresses to clients (options 69,
70).
Installing OS2POPS is simple. Unzip it into the directory where you want it
installed, backup \MPTN\ETC\SENDMAIL.CF, and then run the program
OS2POPS. You will get a screen similar to the one shown in Figure 129 on
page 187.
186
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 129. [OS2POPS] Installation panel
All OS2POPS files should be in the OS2POPS Root Location: (shown on the
screen). After OS2POPS installs, you can configure mail users by clicking on
Client Maintenance, as shown in Figure 130.
Figure 130. [OS2POPS] Client maintenance
When OS2POPS installs, it should modify the C:\MPTN\ETC\SENDMAIL.CF
file, but ours did not. We manually modified the Mlocal entry using the E
editor, as shown in Figure 131 on page 188.
Chapter 4. Serving names
187
Figure 131. [OS2POPS] Manually modifying SENDMAIL.CF
There are single spaces between each entry, not tab characters. Now, the
POP3 mailer is installed and working.
E Editor
Be sure to use the E editor, which will leave tab characters alone. You can
easily break your C:\MPTN\ETC\SENDMAIL.CF file by using an editor that
converts tabs to spaces. Sendmail depends on tabs in certain lines for
proper parsing.
You still need to configure the SMTP mailer. First, start the TCP/IP
Configuration notebook. Then, go to the Autostart tab and set sendmail to
autostart, as shown in Figure 132 on page 189.
188
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 132. [Warp Server] Setting sendmail to autostart
Also notice that the Parameters entry is different than the default. It was
previously SENDMAIL.UML, and it is now SENDMAIL.CF. Configuring a
sendmail file can be very time consuming, and we are not going to describe it
any further than the Mlocal change we did previously. The SENDMAIL.CF file
is as shipped from IBM except for the Mlocal change. Click on OK to save,
and do not reboot at this time (therefore, choose Cancel when asked if you
want to reboot now).
Next, you will update the DHCP server to tell your clients the IP address of
the POP3 and SMTP mailers. Start the DHCP Configuration GUI and
double-click on the global entry, as shown in Figure 133 on page 190.
Chapter 4. Serving names
189
Figure 133. [Warp Server] Modifying global DHCP parameters
Choose the DHCP Options tab, choose option 69 SMTP Server, and Add
the IP address of your server, as shown in Figure 134 on page 191.
190
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 134. [Warp Server] Defining an SMTP server
Then, choose option 70 POP3 Server and Add the IP address, as shown in
Figure 135 on page 192.
Chapter 4. Serving names
191
Figure 135. [Warp Server] Defining a POP3 Server
You see that the mail servers are both on the same machine as the
DHCP/DDNS server, but there is no requirement for this. If you had a
dedicated mail server on our network, you would use that machine’s IP
address instead. Click OK to save your configuration. The DHCP Server
Configuration panel should show the SMTP server and POP3 server options
you configured, as shown in Figure 136 on page 193.
192
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 136. [Warp Server] SMTP and POP3 Global parameters configured
Now, save this configuration. Choose [File => Save] to save the new DHCP
configuration. Re-initialize the DHCP server using the DADMIN -I command.
The next to last task will be updating our DDNS server with an MX record.
Start the DDNS Administration GUI and double-click on the Primary
Domains. Right-click the arrow to show the hidden tabs, and choose the Mail
Exchange tab. As shown in Figure 137 on page 194, choose a preference
number of 10 for your mail server. Note that you use the full canonical name
for your mail server. Because of the way sendmail functions, you are usually
safer using a full canonical name for the mail server. See the O’Reilly
publication, Sendmail, ISBN 1-56592-222-0, for more in-depth information
about sendmail.
Chapter 4. Serving names
193
Figure 137. [Warp Server] Adding a Mail Exchange to the Primary Domain
The preference number is used to direct the mailer to the best server for
performing mail duties. A lower number for the preference value is better, with
zero (0) being the best. The numbers are relative values to each other, not
absolute. It is typical to separate numbers by increments of five or 10 so that
you can insert test mailers without changing any other preference numbers.
When we saved our configuration via the GUI, the domain file was modified,
as in Figure 138 on page 195. You can see the new MX record with value 10.
If we had the dedicated mail server mentioned earlier, we may have entered
two servers, the dedicated one at a value of 10, and our Warp Server at a a
value of 20. This would have caused the dedicated server to be used first,
and OS/2 Warp Server would have been a backup mail server.
194
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 138. [Warp Server] Modification to domain file showing new MX record
The term best server is arbitrary at best, although generally it refers to mail
servers that have a higher capacity or throughput, or a higher bandwidth
connection to where most mail is coming or going. The mailer will attempt to
use lower numbered servers before it attempts to use the higher numbered
mail servers for each mail transaction. If a server is unavailable, extra network
traffic is created. This is one of the reasons to make sure the more reliable
mail servers are numbered lower (of course, user satisfaction is another
reason). The SMTP port number used is 25 and the POP3 port number is
110, as shown in Figure 139 on page 196. This shows the partial output of the
NETSTAT -S command on our OS/2 Warp Server box.
Chapter 4. Serving names
195
Figure 139. [Warp Server] NETSTAT -s showing SMPT and POP3 sockets
Our final task, as with any system change, is to test our new configuration.
We added two users ([email protected] and
[email protected]) to OS2POPS, using the Client
Maintenance as described above. Figure 140 on page 197 shows the
configuration we used on brie, a Windows 95 machine using Netscape
Communicator Version 4.05.
196
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 140. [Netscape] Mail Server configuration
This is representative of Netscape mail setup on any machine. There are
techniques, described in Chapter 7, “Mobile users” on page 317, for
automatically configuring Netscape on client machines. You may be able to
modify the steps for automatic client mail configuration. Figure 141 shows
oyster responding to its first e-mail from brie.
Figure 141. [Netscape] Reading e-mail
Chapter 4. Serving names
197
4.3 Dynamic DNS client support in OS/2 Warp Server
The following sections describe how to enable common clients to work with
Dynamic DNS.
4.3.1 OS/2 Warp 4 with TCP/IP Version 4.1 for OS/2
The following steps show you how to set up the DDNS client on OS/2 Warp 4
using TCP/IP Version 4.1 for OS/2.
4.3.1.1 Enabling DDNS
Follow the steps below to enable the DDNS client:
1. Open the TCP/IP Shadows folder and the TCP/IP Configurations folder,
then double-click on the TCP/IP Configuration (Local) icon, or type
TCPCFG2 at an OS/2 command prompt. The TCP/IP configuration notebook
appears.
2. Click on the Network tab. Select the Automatically, Using DHCP radio
button and check Also, Using DDNS.
198
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 142. [OS/2 Warp] Enabling the DDNS client
3. Click on OK to save the configuration
4. Reboot your computer.
4.3.1.2 Configuring the DDNS client
The DDNS client can be configured using the DDNS Client Configuration
program (DDNSCFG.EXE). You can either type DDNSCFG at an OS/2 command
prompt or double-click on DDNS Configuration in the System Setup folder to
start the program.
When the client is initialized for the first time after DDNS has been enabled,
the DDNS Client Configuration window starts automatically and will be
displayed. Type your hostname. Also type your domain name and DDNS
Chapter 4. Serving names
199
server (they are set by default) if needed. Click Configure to update the
DDNS server.
Figure 143. [OS/2 Warp] DDNS Client Configuration
Starting the next time you start the DDNS Client Configuration program
(DDNSCFG.EXE), previously configured hostnames are read-in from the
keyfile and displayed for the user to select. Figure 144 on page 201 shows an
example.
200
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 144. [OS/2 Warp] Selecting a previous configuration
Once registered, the DHCP Client Monitor displays the hostname. To check
the hostname, double-click on the DHCP Monitor object in the System Setup
folder.
4.3.2 OS/2 Warp 4 with TCP/IP 4.0
The following steps show you how to set up a DDNS client on OS/2 Warp 4
using the TCP/IP version that comes with the product, which is TCP/IP V4.0.
4.3.2.1 Enabling DDNS
Follow the steps below to enable the DDNS client:
1. Open the Programs folder and TCP/IP Internet (LAN) and then
double-click on TCP/IP Configuration (LAN), or type TCPCFG at an OS/2
command prompt. The TCP/IP configuration notebook appears.
2. Select the radio button that says Automatically, Using DHCP and mark
the Also, Using DDNS check box.
Chapter 4. Serving names
201
Figure 145. [OS/2 Warp] Enabling the DDNS client
3. Exit the Configuration notebook by closing the window and saving the
changes.
4. Reboot your computer.
4.3.2.2 Configuring the DDNS client
The DDNS client can be configured using the DDNS Client Configuration
program. You can either type DDNSCFG at an OS/2 command prompt or
double-click on DDNS Configuration in the System Setup folder to start the
program.
When the client is initialized for the first time after DDNS has been enabled,
the DDNS Client Configuration window starts automatically. Type your
hostname. Also, type your domain name and DDNS server (they are set by
default), if needed. Click Configure to update the DDNS server.
202
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 146. [OS/2 Warp] DDNS Client Configuration
Once registered, the DHCP Client Monitor displays the hostname.
Double-click on the DHCP Monitor icon in the System Setup folder.
Figure 147. [OS/2 Warp] DHCP Monitor
4.3.3 OS/2 Warp 4 using presecured mode
This section explains how to configure an OS/2 Warp 4 DDNS client using the
dynamic presecured domain environment.
4.3.3.1 Enabling DDNS
Follow the steps below to enable a DDNS client:
Chapter 4. Serving names
203
1. Open TCP/IP Shadows folder and TCP/IP Configurations folder, then
double-click on the TCP/IP Configuration (Local) icon, or just type
TCPCFG2 at an OS/2 command prompt. The TCP/IP Configuration Notebook
appears.
2. Click on the Network tab. Select the Automatically, Using DHCP radio
button and check Also, Using DDNS.
Figure 148. [OS/2 Warp] Enabling the DDNS client
3. Click OK to save the configuration.
4. Reboot your computer.
4.3.3.2 Configuring DDNS client
The DDNS client can be configured using the DDNS Client Configuration
program. You can either type DDNSCFG at an OS/2 command prompt or
204
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
double-click on the DDNS Configuration object in the System Setup folder to
start the program.
When the client is initialized for the first time after DDNS has been enabled,
the DDNS Client Configuration starts automatically and its window appears.
Click Import Key File to import the key file to the local machine. Change to
the directory where key file is stored, choose the appropriate key file and click
OK.
Figure 149. [OS/2 Warp] Importing key file
If the DDNS.DAT file already exists on the local machine, a panel, as shown in
Figure 150, appears. You can either replace or append to the existing file.
Figure 150. [OS/2 Warp] DDNS Configuration Message
Chapter 4. Serving names
205
After copying the key file, the hostname, domain name, and Dynamic DNS
server name are automatically filled in the DDNS Configuration notebook.
Click Configure to update the DDNS server.
Figure 151. [OS/2 Warp] DDNS Client Configuration
Once registered, the DHCP Client Monitor displays the hostname. To check
the hostname, double-click on the DHCP Monitor icon in the System Setup
folder.
4.3.4 WorkSpace On-Demand using proxy
Since WorkSpace On-Demand client loads all the software it needs from
WorkSpace On-Demand Manager (server), all the configuration is performed
at the server.
4.3.4.1 Configuration
To enable the DDNS client for a particular WorkSpace On-Demand system,
follow the steps below. Also, see 2.5.11, “IBM WorkSpace On-Demand” on
page 102.
206
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
1. Start the LAN Server Administration program and open the Remote IPL
Requesters folder.
2. Press the WorkSpace On-Demand Client icon.
3. Make sure that DHCP client is enabled by checking Automatic, using
DHCP.
4. Check Also using DDNS. By doing this, fields other than the hostname and
domain name are grayed-out. Fill in the hostname and domain name. Click
Set to save and close the notebook.
Figure 152. [WSOD] TCP/IP Setting notebook
The hostname you specified in the WSOD client notebook is added as Option
12 (hostname) and the domain name is added as Option 15 in the DHCP
Client Configuration file (DHCPCD.CFG). These are sent to a DHCP server.
Start the WSOD client. The WSOD client gets an IP address from the DHCP
server and the DHCP server registers the hostname for the WSOD client.
In a regular OS/2 Warp machine, the DDNS Client Configuration program
(DDNSCFG.EXE) is launched automatically when the client is initialized for
the first time after DDNS has been enabled. The DDNS client starts
automatically every time the machine is initialized. However, this does not
Chapter 4. Serving names
207
occur at the WSOD client. Even though you check Also using DDNS in the
TCP/IP configuration, the WSOD client does not register its hostname by
itself, and a ProxyArec-enabled DHCP server would update the DDNS server.
Tip: Delete DDNSCFG.EXE from the RIPL tree
If you open a DHCP monitor on the WSOD client, you will notice that it says
Registration Failed, as shown in Figure 153. This is OK. The reason is that
the monitor only shows a hostname registered by the DDNS Client
Configuration program. (DDNSCFG.EXE). It does not show a hostname
registration through ProxyArec.
We recommend administrators deleting DDNSCFG.EXE from the RIPL tree
so that this program cannot be accidentially executed by the WorkSpace
On-Demand user.
Figure 153. [WSOD] DHCP Client Monitor
4.4 Name serving on AIX
The following sections will provide information on configuring AIX 4.3.3 as a
DNS client, static DNS server, and dynamic DNS server. For AIX DNS client
configuration, the file set bos.net.tcp.client will have to be installed. For AIX
DNS server services, bos.net.tcp.server will have to be installed. Both of
these file sets are available on your AIX install media. Note that the
configuration of AIX as a static DNS is a prerequisite for configuring AIX for
DDNS, since DDNS configuration is built on top of existing static DNS
services.
208
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
4.4.1 AIX as a DNS client
To have an AIX server or workstation utilize a name server on the network,
type smitty at the command line, choose Communication Applications and
Services, then TCP/IP, then Further Configuration, then Name Resolution,
and you will be prompted with the screen as shown in Figure 154 on page
209. Highlight Start Using the Nameserver and press Enter.
Figure 154. [AIX DNS] Client configuration - Use nameserver
As seen in Figure 155 on page 210, the first field will contain the IP address
of the name server you wish to point your host towards, and the second field
will contain the domain name on which your host resides. When both of the
fields are completed, press Enter.
Chapter 4. Serving names
209
Figure 155. [AIX DNS] Client configuration - Specifying the nameserver
After the command has completed successfully, press F3 twice to exit, and
you will be returned to the Domain Nameserver smit screen. As shown in
Figure 156 on page 210, move the cursor down to Set / Show the Domain
Search List and press Enter.
Figure 156. [AIX DNS] Client configuration - Domain search list
210
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 157 shows the domain armonk.cooking.net added to the domain
search list. After this is entered, press Enter and the entry will be added to
/etc/resolv.conf.
Figure 157. [AIX DNS] Client configuration entering the search domain
4.4.2 AIX as a static DNS server
Information on configuring AIX 4.3.3 as a DNS server can be found at:
http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/aixbman/commadmn/tcp_na
meres.htm#BCAF2DF137dada
The instructions provided at this site help with creating the configuration files
necessary for the named daemon to function properly. If the fileset
bos.net.tcp.server is installed you will be able to find sample versions of all
the DNS configuration files within the directory /usr/samples/tcpip. Also, we
have put copies of our named .* configuration files on the CD-ROM that
comes with this book. On the CD-ROM, they are located in the directory
\AIX\DNS.
The first file, named.boot can be seen in Figure 158 on page 212. It shows
that forward name resolution definitions for the domain are in the file
named.forward, and the reverse mapping definitions are in named.rev. There
is also a definition for local name resolution in named.local. Notice that for the
reverse and local zone definitions, the domain format is different. The IP
address for the name server is 192.168.6.4, so the domain name for the
reverse mapping zone is 6.168.192.in-addr.arpa. Similarly, the IP address for
Chapter 4. Serving names
211
loopback is 127.0.0.1, so the domain name for local name resolution
becomes 0.0.127.in-addr.arpa.
Figure 158. [AIX DNS] Server configuration - named.boot
Figure 159 on page 213 shows the named.conf file created in the /etc
directory. Here, designate where all of your configuration files are going to be
with the directory field, and then define your zone for forward name
resolution, reverse name resolution, and then a zone for local name
resolution.
Note
The files named.boot and named.conf are mutually exclusive in DNS BIND
8.2.2, so either one or the other should be used, but not both. See 4.4.3.1,
“Updating the BIND level” on page 218 for updating to BIND 8.2.2.
212
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 159. [AIX DNS] Server configuration - named.conf
As can be seen in the named.forward file in Figure 160, we have our main
definition for the SOA (Start Of Authority) record, the definition for our NS
(Name Server) resource record, and then at the bottom, the A resource
record definitions for our name server (buffet) and for other hosts on the
network. For an explanation of A, SOA, and NS records, please see 1.6.5,
“Resource record types” on page 23.
Figure 160. [AIX DNS] Server configuration - named.forward
Chapter 4. Serving names
213
Furthermore, in Figure 161 on page 214 in our named.rev file, we are making
declarations for our SOA record for reverse mapping and our PTR record for
our name server. At the bottom of the file are PTR resource record definitions
for other hosts on our network. Our domain definition in named.conf for
reverse mapping in Figure 159 on page 213 was 6.168.192.in-addr.arpa,
making the PTR record of 2 IN PTR mustard.armonk.cooking.net. resolve to
2.6.168.192.in-addr.apra for reverse lookup. Note that at the end of each fully
qualified host name there is a period, like mustard.armonk.cooking.net. and
for router.armonk.cooking.net. For an explanation of PTR resource records
please see 1.6.5, “Resource record types” on page 23.
Figure 161. [AIX DNS] Server configuration - named.rev
Figure 162 on page 215 shows the named.local file that has the SOA
definition for local name resolution, NS (Name Server) record for local name
resolution, and finally the PTR record for the loopback address pointing to the
local host. As shown in Figure 159 on page 213, our definition for reverse
mapping of our local domain was 0.0.127.in-addr.apra, so that the PTR
definition of 1 would have a fully qualified domain name for reverse mapping
of 1.0.0.127.in-addr.arpa.
214
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 162. [AIX DNS] Server configuration - named.local
As shown in Figure 163 on page 215, edit /etc/resolv.conf and add the IP
address for your host as the nameserver and domain. This is so that when
named starts, it knows to look back upon itself for name resolution.
Figure 163. [AIX DNS] Server configuration - /etc/resolv.conf
Chapter 4. Serving names
215
After these files are configured, you can enter the command startsrc -s
named from the command line to start the named subsystem and enable the
RS/6000 as a DNS server.
4.4.2.1 Using awk scripts to create DNS zone files
As an alternative to creating the AIX DNS zone files manually, you can use
awk scripts to create the named.forward and named.rev files. The steps
below describe how to do this:
1. Make sure your DNS hostname is set to the fully-qualified name. For
instance, our DNS host is called buffet on domain armonk.cooking.net.
Typing hostname on this system should return buffet.armonk.cooking.net. If
it does not, type smit hostname to set the hostname to the fully-qualified
name. This ensures that the domain information in the zone files is
complete.
2. Create an /etc/hosts file on your DNS server. We will use the following
information to create the initial hosts file:
192.168.6.1
192.168.6.2
192.168.6.10
192.168.6.15
router
mustard
merlot
buffalo
3. Now, go to the directory that contains the awk scripts. Type:
cd /usr/samples/tcpip
4. Next, run the awk scripts. Run one script to create the named.forward file,
and a second script to create the named.rev file:
./hosts.awk /etc/hosts > /etc/named.forward
./addrs.awk /etc/hosts > /etc/named.rev
The hosts.awk script creates a zone file with the A records for our entries
in the hosts file we specified above. The addrs.awk file creates the PTR
records for these same entries in the hosts file.
Make sure to erase the /etc/hosts file (or at least rename it) to avoid the
chance that it might be referenced when the DNS server is resolving names.
Figure 164 on page 217 shows the named.forward file that results from
running the hosts.awk script, and Figure 165 on page 217 shows the
216
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
named.rev file created using the addrs.awk script.
; name server data file
; (also see /etc/named.boot)
;
; NAME
TTL
CLASS TYPE RDATA
;
; setting default domain to "armonk.cooking.net"
;
@
9999999 IN
SOA
buffet.armonk.cooking.net.
root.buffet.armonk.cooking.net.(
1.1
; Serial
3600
; Refresh
300
; Retry
3600000 ; Expire
86400 ) ; Minimum
9999999 IN
NS
buffet
router
9999999 IN
A
192.168.6.1
mustard 9999999 IN
A
192.168.6.2
merlot
9999999 IN
A
192.168.6.10
buffalo 9999999 IN
A
192.168.6.15
Figure 164. [AIX DNS] Named.forward file created by hosts.awk
; setting default domain to ... armonk.cooking.net
@
9999999 IN
SOA
buffet.armonk.cooking.net.
root.buffet.armonk.cooking.net. (
1.1
;Serial
3600
;Refresh
300
;Retry
3600000
;Expire
86400 )
;Minimum
9999999 IN
NS
buffet.armonk.cooking.net.
1.6.168.192
IN PTR router.armonk.cooking.net.
2.6.168.192
IN PTR mustard.armonk.cooking.net.
10.6.168.192
IN PTR merlot.armonk.cooking.net.
15.6.168.192
IN PTR buffalo.armonk.cooking.net.
Figure 165. [AIX DNS] Named.rev file created by addrs.awk
Chapter 4. Serving names
217
4.4.3 AIX as a DDNS server
Note
This section has the following prerequisites:
1. Proper configuration of AIX as a DHCP server. This is covered in 2.4,
“AIX as a DHCP server” on page 50.
2. Proper configuration of AIX as a static DNS server. This is covered in
4.4.2, “AIX as a static DNS server” on page 211.
In this section we discuss two DDNS scenarios on AIX. First, we will cover the
configuration of DDNS allowing unsecured zone updates along with
ProxyArec so that our AIX DDNS server handles the updating of both PTR
and A resource record updates for our DHCP clients. However, configuring
these zones in unsecured mode leaves our DDNS server open to IP spoofing.
IP spoofing is when a hacker maps their IP address to pretend to be another
device on your network.
To prevent a hacker from pretending to be our DDNS server and doing any A
and PTR resource record updates, configure your DDNS in controlled mode.
Putting your DDNS in controlled mode requires an encryption key from your
DHCP server in order to perform A and PTR resource record updates.
4.4.3.1 Updating the BIND level
The default installation of bos.net.tcp.server 4.3.3.0 installs the BIND 4.9.3
version of the named daemon and it has been reported that this version of
named has issues with taking updates from nsupdate to dynamically update
zones. Therefore, your first step will be to bring the name server daemon up
to BIND Version 8.2.2.
As can be seen in Figure 166 on page 219, you are in the /usr/sbin directory
where the binary files for the named daemon are kept. Remove the existing
named and named-xfer files and replace them by linking them with the
named8 and named8-xfer binaries. This can be done with the following
commands:
rm
rm
ln
ln
218
named
named-xfer
-s named8 named
-s named8-xfer named-xfer
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 166. [AIX DDNS] Unsecured server configuration - LN command
4.4.3.2 AIX DDNS in unsecured mode with ProxyArec
Note
The configuration files used in this section are included in the CD that
comes with this book. They are located in the path /AIX/DDNS/unsecured.
Next, change /etc/dhcpsd.cnf to allow the dhcpsd daemon to make updates to
the DNS server via /usr/sbin/dhcpaction. As can be seen in Figure 167 on
page 220, you should add the line:
updateDNS "/usr/sbin/dhcpaction '%s' '%s%' '%s' '%s' BOTH NONIM >>
/tmp/updns.out 2>&1".
The field where you enter BOTH can be either BOTH, A, or PTR. By setting
the field to BOTH, you are configuring your DHCP server to update both the A
and PTR records for your DHCP client, essentially enabling ProxyArec.
Chapter 4. Serving names
219
Figure 167. [AIX DDNS] Unsecured server configuration - dhcpsd.cnf
Figure 168 on page 221 shows the necessary changes to the named.conf file
for DDNS on AIX.
The first change in both zones is the addition of the allow-update field, where
you designate which host address is allowed to make updates to the zone.
Here, you specify the IP address of your name server, since the scenario is
going to be unsecured with ProxyArec.
Second, update-security specifies whether or not the DNS server must pass
an encryption key to nsupdate to update the zone. Here, you specify
unsecured so that no encryption key will be required. There are two other
settings available: presecured and controlled. The presecured setting will
require a private key from the client and a public key from the server, while
controlled only requires a public key from the server.
Third, incr-interval specifies an interval in which the name server will accept
dynamic updates while not increasing the serial level of the zone's SOA
record. Because a change in the zone SOA record will instantiate a NOTIFY
message, limiting this occurrence will minimize the number of zone transfer
requests at the expense of minimal zone differences between a dynamic
master server and its slave.
220
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Fourth, deferupdcnt specifies a threshold value for the number of properly
applied updates received during one incr-interval interval. If more than this
value of updates are realized during the interval, the name server will modify
the zone SOA serial level and subsequently notify each of the slave servers.
Use this value to limit the database replication inconsistencies in an
environment where dynamic zone updates occur infrequently but in large
quantities.
More can be read about these fields for named.conf at:
http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/files/aixfiles/named.co
nf.htm
Figure 168. [AIX DDNS] Unsecured server configuration - named.conf
Once these changes are completed, you will want to stop the named
subsystem with the command stopsrc -s named and then restart it with the
command startsrc -s named. Next, stop the dhcpsd subsystem with the
command stopsrc -s dhcpsd and restart it with startsrc -s dhcpsd to
implement the changes made to /etc/dhcpsd.cnf. You can view
/tmp/updns.out to read the DDNS log and to view the updates that have taken
place.
Chapter 4. Serving names
221
4.4.3.3 AIX DDNS in controlled mode with ProxyArec
Note
The configuration files used in this section are located on the CD-ROM that
came with this redbook in the directory /AIX/DDNS/controlled.
The first step is to stop the named and dhcpsd subsystems. Do this using the
stopsrc command, as illustrated in 4.4.3.2, “AIX DDNS in unsecured mode
with ProxyArec” on page 219.
To configure our AIX DDNS server to function in controlled instead of
unsecured mode, change the update-security field in named.conf from
unsecured to controlled in both forward and reverse lookup zones. In Figure
169 on page 222, you can see the changes made to /etc/named.conf.
Figure 169. [AIX DDNS] Controlled server configuration - named.conf
Next, generate the encryption key to be handled by the zone and passed to
the zones for updates by using the nsupdate command. This will be done with
the following command:
222
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
nsupdate -g -k /etc/zonekey -h armonk.cooking.net -p
buffet.armonk.cooking.net
After the key is generated, copy the public key (the latter part of the key file)
and paste it into the SOA records in named.forward and named.rev. In Figure
170 on page 223, you can see the highlighted public part of the key. Once this
is highlighted with the left mouse button, hold down the Ctrl key and press
Insert. This copies the contents of the highlighted text, so you are able to
paste the key into text.
Figure 170. [AIX DDNS] Controlled server configuration - Public Key
In Figure 171 on page 224, you can see where, in named.forward, the KEY
resource record definition has been added for the zone armonk.cooking.net
with the line:
INKEY0x0100 0 1 public key
To insert the public key info, copy from /etc/zonekey by holding down the Shift
key and pressing Insert. For an explanation of the KEY resource record, see
1.6.5, “Resource record types” on page 23.
Chapter 4. Serving names
223
Figure 171. [AIX DDNS] Controlled server configuration - named.forward
In Figure 172 on page 224, you can see that the same steps are repeated in
named.rev as were performed on named.forward.
Figure 172. [AIX DDNS] Controlled server configuration - named.rev
Finally, you will edit the calls to nsupdate that are made by
/usr/sbin/dhcpaction. There are two occurrences of the nsupdate command in
the file you must edit. In Figure 173 on page 225, you can see where the flag
224
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
-k /etc/zonekey was inserted into the call to nsupdate. There is another
occurrence of the nsupdate command later in the dhcpaction script that will
have to be changed as well.
Figure 173. [AIX DDNS] Controlled server configuration - /usr/sbin/dhcpaction
Now, you can restart the subsystems using the startsrc command as
discussed in 4.4.3.2, “AIX DDNS in unsecured mode with ProxyArec” on page
219. The named daemon will now accept dynamic updates from dhcpsd in
controlled mode.
Chapter 4. Serving names
225
226
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 5. Integrating file and print services
In this chapter, we show how to integrate your existing LANs into your TCP/IP
network. This chapter focuses on the legacy protocol NetBEUI, NetBIOS
name servers, Microsoft WINS, and how to enable your clients to use these
services. If you are interested in integrating printing and file sharing into your
TCP/IP network, then you should read this chapter.
5.1 NetBIOS, NetBEUI and TCPBEUI: What are they?
NetBIOS (Network Basic Input Output System) was originally developed by
Sytek Inc. for IBM to link a network operating system with specific hardware.
NetBIOS offers LAN applications a variety of hooks to carry out
inter-application communications and data transfer. Essentially, NetBIOS is a
way for application programs to talk to the network.
NetBEUI (NetBIOS Enhanced User Interface) is the original PC networking
protocol and interface designed by IBM for OS/2 LAN Server/OS/2 Warp
Server Version 4. This protocol was later adopted by Microsoft for their
networking products.
NetBEUI uses the IEEE standard 802.2 data link protocol. In addition to
connectionless communication, this protocol offers connection-oriented,
guaranteed delivery of messages up to 64 Kb in length. The
connection-oriented flavor of 802.2 exploits advanced windowing and
acknowledgment efficiencies, and renders stellar performance when both
ends of the connection are on the same LAN. When the connection spans
multiple LANs, the 802.2 protocol relies on Source Route Bridging or
Transparent Bridging. NetBEUI can not be routed by an IP gateway.
Fortunately, the NetBIOS interface does not rely on any specific underlying
protocol to connect systems together, and as such, it is adaptable to
protocols that allow packet forwarding, such as TCP/IP.
Because NetBEUI is not a routable protocol, it is a major limitation for a
growing company where routable protocols, such as TCP/IP, are gaining
popularity.
RFCs 1001 and 1002 describe the standard way to implement the NetBIOS
services on top of the Transmission Control Protocol (TCP) and User
Datagram Protocol (UDP). RFC 1001/1002 is not an encapsulation technique;
it builds special packets and sends them out via UDP/IP and TCP/IP. NetBIOS
over TCP/IP is known as NBT or TCPBEUI.
© Copyright IBM Corp. 1998, 2000
227
Note
TCPBEUI is translation, not encapsulation.
Once a NetBIOS session has been established, TCPBEUI will use
sockets-send commands over a TCP connection to send NetBIOS session
data. TCPBEUI builds a 4 byte session header that precedes the actual
user data. Therefore, a NetBIOS chain send of 128 KB would have an
overhead of only 4 bytes.
The IBM OS/2 Warp Server for e-business and Microsoft Windows NT
implement RFCs 1001 and 1002 in order to overcome the limitation of
NetBEUI. Figure 174 on page 229 shows how a workstation can
communicate across an IP-only router with traditional TCP/IP hosts, as well
as with servers using the TCPBEUI protocol. The workstation can continue to
communicate with local servers using NetBEUI.
228
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 174. Workstation communicating using multiple protocols
Figure 175 on page 230 shows the structure of NetBIOS, NetBEUI and
TCPBEUI when implemented under OS/2. Data flows within the diagram are
vertical only.
Chapter 5. Integrating file and print services
229
Figure 175. [OS/2 Warp] NetBIOS, NetBEUI and TCPBEUI structure
Whenever a NetBIOS-based computer (NetBIOS, NetBEUI or TCPBEUI)
connects to a network, it registers several NetBIOS names to the network.
NetBIOS names are made up of 16 bytes that normally consist of a 15
characters name with a one character suffix. All 16 bytes can be binary.
These names can be unique to the workstation (for example, the
computername), or they can be a group name (for example the workgroup or
domain name). Table 9 on page 231 lists the various names and suffixes
used.
230
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Table 9. NetBIOS names and suffixes
Name
Number
Type2
Usage
1
computername
00
U
Workstation service.
computername
01
U
Messenger service.
computername
03
U
Messenger service used when sending and receiving
messages. This is the name registered with the WINS
server.
computername
05
U
Messenger Service. It is used to check if the name
registered has been forwarded to another workstation.
computername
06
U
RAS server service.
domain
1B
U
Domain master browser, which clients use to contact the
domain master browser.
domain
1D
U
Master browser for server lists.
computername
1F
U
NetDDE service.
computername
20
U
File server service.
computername
21
U
RAS client service.
computername
BE
U
Network Monitor Agent.
computername
BF
U
Network Monitor Application.
domain
00
G
Domain name.
__MSBROWSE__
01
G
The name master browsers broadcast to announce their
domains to other Master Browsers.
domain
1C
G
Registered by the Domain Controller, it contains a list of
specific addresses of systems that have registered the
name.
domain
1E
G
Browsers broadcast and then listen on this name to elect a
master browser. This broadcast should be on the local
subnet only.
groupname
20
G
A special name registered with WINS servers to identify
groups of computers for management purposes.
1 Numbers are hexadecimal
2 U = Unique, G = Group
Chapter 5. Integrating file and print services
231
To communicate over TCPBEUI, two computers refer to one-another using
these names. A problem occurs because they must somehow resolve these
NetBIOS names into IP addresses.
5.2 Resolving NetBIOS names to IP addresses
When the client connects to the server the client must resolve the server’s
NetBIOS names. There are various ways to resolve NetBIOS names into IP
addresses.
The simplest method is to try broadcasting name queries and hope the
computer you’re trying to communicate with responds in order to do the
resolution. Computers running in this mode are said to be B-Node
(Broadcast) clients.
A better method would be to have a server on the network that could do the
NetBIOS name to IP addresses resolution for you. Such a server is known as
a NetBIOS Name Server or NBNS. Microsoft uses WINS as a type of NBNS.
Computers that are configured to use a NBNS are P-Node (Peer-to-Peer)
clients. Broadcasting is never used by P-Node clients.
Finally, you could combine the modes to become either M-Node (Mixed) or
H-Node (Hybrid) clients.
An M-Node client will attempt a name query broadcast first, and if that fails, it
will try to use the NBNS.
H-Node clients reverse the sequence used by M-Node clients, and will
attempt to use the NBNS first. If that fails, it will try a name query broadcast.
If an H-Node client detects that a NBNS has failed, it will continue to poll the
NBNS (while using broadcasts), so that it knows when to switch back to using
the NBNS. H-Node has generally replaced M-Node.
NetBIOS name servers will be covered in 5.4, “NetBIOS Name Servers
(NBNS)” on page 253. Configuring clients to make use of the NBNS is
covered in 5.5, “P-Node, H-Node and M-Node clients” on page 257. In the
meantime, we will look at how to enhance the performance of B-Node clients
by reducing the amount of broadcasts that they generate.
There is a new way of resolving a NetBIOS name to IP address, not using a
broadcast or NetBIOS name server. This approach is discussed in 5.3, “New
way of using DNS with non-RFC-encoded name” on page 246.
232
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
5.2.1 Enhancing B-Node clients
Operating as a B-Node client will work for you as long as all the routers and
bridges within your network are configured to pass on broadcast traffic.
However, generally they’re not, so computers will only be able to
communicate with one another if they are located on the same subnet.
There are enhancements available to B-Node clients in order to work around
this limitation.
5.2.1.1 Enhancing DOS, Windows 3.11 and WfW clients
The IBM DOS LAN Services (DLS) comes with a TCP/IP stack from Network
TeleSystems that supports DHCP and TCPBEUI. Throughout this chapter,
when we’re discussing DOS, Windows 3.11 or Windows for Workgroups
(WfW)-based computers, we will assume that DLS is installed.
DLS also includes the PING utility for testing IP connectivity, and a utility
called NBUTIL to build a NetBIOS name to IP address resolution table.
The NBUTIL Utility Program
The NBUTIL utility program that comes with DLS allows you to create and
maintain a NetBIOS name to IP address table in memory.
NBUTIL syntax
NBUTIL [-?]
[-a <ipaddress> <netbiosname> ]
[-f <filename>]
[-l <number>]
[-x]
[-t]
[-s]
[-c]
[-w]
Where:
?
Displays a description of all command line options used with
NBUTIL.
-a
Adds a NetBIOS name/IP address pair to the NetBIOS
name table.
<ipaddress>
The IP address associated with the NetBIOS name you want
to add to the NetBIOS name table.
Chapter 5. Integrating file and print services
233
<netbiosname> Is the NetBIOS name for the remote system you want to add
to the NetBIOS name table.
-c
Clears all entries from the NetBIOS name table.
-f
Specifies the ASCII file containing the NetBIOS names with
their associated addresses.
<filename>
The name of an ASCII text file containing the NetBIOS
names with their associated IP addresses.
-l <number>
This option represents the LAN adapter number
(LANABase) used by the TCP/IP protocol driver. If -l is not
specified, a default of (0) is used.
-s
Displays all the current entries in the NetBIOS name table.
-t
Enables you to put embedded spaces in NetBios names by
translating an underscore(_) into a space. When used with
the -s option, -t translates spaces back into underscores.
-w
Translates NetBIOS names into the format of the LAN
manager workstation name.
-x
Translates the NetBIOS name you want to add to the
NetBIOS name table into names that can be used by
servers running LAN manager.
Note
The NetBIOS names table is for NetBIOS to IP address resolution only. It is
not used for IP host name to IP address resolution.
Using NBUTIL
The NBUTIL command must be executed after the TCP/IP protocol driver is
running; that is, after DLS has started.
You can use the NBUTIL command to add entries to the NetBIOS name table in
two different ways:
• By adding the IP address of each system you want to access with its
associated NetBIOS name one at a time
• By creating an ASCII text file containing a number of IP addresses with
their associated names, and by using the -f option, all of these names can
be added to the NetBIOS name table by invoking the NBUTIL command only
once
234
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The format of the file is:
<IP address> <NetBIOS name> <# comment>
<IP address>
Is the IP address of the remote system you want to
access
<NetBIOS name>
Is a NetBIOS name associated with the IP address of
the remote system you want to access
<#comment>
Is a number symbol (#) followed by any additional
information you want to add to the file as a comment;
the # comment is optional
192.168.6.10 merlot # The MERLOT server
192.168.7.10 texmex # The TEXMEX server
Figure 176. [DOS LAN Services] Example input file for NBUTIL
If you are running DLS under DOS and want to run the command
automatically each time your PC is started, place the command in the
AUTOEXEC.BAT file after the NET START command.
@ECHO OFF
SET PATH=C:\;C:\NET;C:\DOS
PROMPT $P$G
LH MOUSE
LH SHARE
SET TCPHELP=C:\NET
SET ETCDIR=C:\ETC
C:\NET\NET START
NBUTIL -A 192.168.7.10 texmex -X
Figure 177. [DOS LAN Services] Example AUTOEXEC.BAT using NBUTIL
If you are running DLS under Windows, you must run the NBUTIL program
from a DOS box after Windows has been started. This is because the protect
mode driver does not load until Windows has started. You must wait until then
to run NBUTIL. This can be done automatically by creating an icon to the
Startup folder.
Another option to consider is using LAN logon scripts to centrally manage the
names that are in the table.
As a test, we created a user ID on a Warp Server for e-business domain, that
was assigned a Z: drive at logon that contained a file called NBUTIL.LST.
Chapter 5. Integrating file and print services
235
(Identical to the example file shown in Figure 176 on page 235). The user’s
PROFILE.BAT contained the command:
C:\NET\NBUTIL -F Z:\NBUTIL.LST -X
After logging onto the LAN from either a DOS DLS or a Windows DLS
workstation, the NBUTIL -S command was run to produce the output shown in
Figure 178
C:\>NBUTIL -S
NBUTIL - Netbios Name Utility 2.09 (960320)
Copyright (c) 1994-1996 Network TeleSystems, Inc. All rights reserved.
There are 8 entries in the Name/IP cache:
TEXMEX
TEXMEX
TEXMEX
TEXMEX
MERLOT
MERLOT
MERLOT
MERLOT
192.168.7.10
© 192.168.7.10
192.168.7.10
ß 192.168.7.10
192.168.6.10
© 192.168.6.10
192.168.6.10
ß 192.168.6.10
Figure 178. [DOS LAN Services] Sample output from NBUTIL
You could also load the file using a UNC name. For example:
C:\NET\NBUTIL -F \\MERLOT\CONFIGS\NBUTIL.LST -X
5.2.1.2 Enhancing Windows 95, Windows 98 & Windows NT clients
Windows 95, Windows 98 and Windows NT can all use a file called
LMHOSTS to perform NetBIOS name to IP address resolution.
Under Windows 95 and Windows 98, the file should be located in the
\WINDOWS subdirectory, while under NT, it is located in
\WINNT\SYSTEM32\DRIVERS\ETC. If the file does not exist, a sample file
called LMHOSTS.SAM in these subdirectories can be copied to LMHOSTS.
At its simplest, the file contains the NetBIOS name and IP address pairs.
Additional options are available to enhance the configuration.
LMHOSTS options
All options should be in upper case.
# <comment text>
236
Indicates that any text following is a comment unless it
is one of the options. # is used with all the keywords to
maintain backwards compatibility with LAN Manager,
which will see the options as comments.
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
#PRE
Specifies that the address should be pre-loaded into
the name cache. The LMHOSTS file is used only after
a WINS query and a broadcast have both failed, and
these searches are attempted only if the name is not
already in the cache.
#DOM:<domain name>
This option is needed for all servers that validate
logon requests. All computers in the domain will need
one of these entries for each domain controller in the
domain. Note that #PRE must be before #DOM.
#NOFNR
Do not use Directed Name Queries (DNS). Some
versions of LAN Manager cannot handle NetBIOS
DNS requests and return an error.
#BEGIN_INCLUDE
#INCLUDE <other_file>
#INCLUDE <other_file>
#END_INCLUDE
Include LMHOSTS definitions from other files. The
other file may be remote. (Remember, the computer
must be able to resolve the name of the remote
computer before it can use a remote file.)
#BEGIN_ALTERNATE
#INCLUDE <other_file>
#INCLUDE <other_file>
#END_ALTERNATE
Defines a list of alternate locations for an additional
LMHOSTS file to be included. The first additional file
read successfully will be used.
Associates a NetBIOS name with multiple IP
addresses (For example, a server with two or more
NICs). The NetBIOS name can be associated with up
to a maximum of 25 different IP addresses.
#MH
Note
The name cache and the LMHOSTS file are read sequentially, so add the
most frequently accessed computer names to the top of the LMHOSTS file,
with the #PRE entries at the end of the file. Once entries have been loaded
into the cache, they do not need be accessed again in the file.
Figure 179 on page 238 shows a sample LMHOSTS file for a computer that
can participate in two domains (REDMOND and BELLVUE). Using the
#BEGIN_ALTERNATE option, an additional LMHOSTS configuration file is
loaded from either the PIE or HOTDOG server (The first one successfully
Chapter 5. Integrating file and print services
237
loaded will be used). An entry also exists for a DOS-based workstation called
BURGER.
Notice that the server entries are located before the #BEGIN_ALTERNATE
entries and they are pre-loaded. This allows the server names in the alternate
entries to be successfully resolved without having to refer back to the
LMHOSTS file.
192.168.6.13 BURGER # DOS workstation
192.168.8.10 PIE
#DOM:REDMOND #PRE # PDC for REDMOND domain
192.168.7.9 HOTDOG #DOM:BELLVUE #PRE # PDC for BELLVUE domain
#BEGIN_ALTERNATE # Load further configuration file
#INCLUDE \\PIE\CONFIGS\LMHOSTS
#INCLUDE \\HOTDOG\CONFIGS\LMHOSTS
#END_ALTERNATE
Figure 179. [Windows 95/NT] Sample LMHOSTS file
With Windows 95, the LMHOSTS file must exist to be used by the operating
system. With Windows NT, use of LMHOSTS must be explicitly turned on by
selecting the Enable LMHOSTS Lookup checkbox as shown in Figure 180
Figure 180. [Windows NT] Enabling use of LMHOSTS
238
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Windows 95 computers may encounter problems when trying to NET VIEW a
domain that is on the other side of a router if the computer is only using an
LMHOSTS file to do NetBIOS names resolution and not a NBNS. It is
necessary to add the following to the LMHOSTS file.
<IP address> <tab> "<domain_name>
\0x1B" <tab> #PRE
<IP address>
The IP address of the primary domain controller.
<tab>
A TAB character.
<domain_name>
The name of the domain. If the domain name is less than
15 characters, use spaces to pad it to 15 characters. The
hexadecimal value 1B is appended following the Domain
name as the 16th character (a total of 20 characters
including the "\0x1B"). The Domain name must be all
upper-case letters, and must be inside quotation marks.
5.2.1.3 Enhancing OS/2 Warp clients
Under OS/2 Warp, there are three enhancements available to reduce the
amount of name query broadcasts generated by computers using TCPBEUI.
Two of these require configuring text files on each computer, while the other
makes use of an existing DNS on the network.
Names file routing extension
This routing extension is implemented through the use of a names file that
contains NetBIOS name and IP address pairs. The names file, called
RFCNAMES.LST, is located in the \IBMCOM directory.
This implementation requires an entry in the names file for every workstation
with which you want to communicate. This file is conceptually similar to the
TCP/IP hosts file.
Note
The RFCNAMES.LST file is for NetBIOS name to IP address resolution
only. It is not used for IP host name to IP address resolution.
The names file routing extension is enabled by setting the NAMESFILE
parameter in the PROTOCOL.INI file equal to a nonzero integer value, which
represents the maximum number of entries in the names file.
The RFCNAMES.LST format is:
<string> <IP address>
Chapter 5. Integrating file and print services
239
Where:
<string>
A string of characters used when searching for NetBIOS
names in the file, delimited by double quotation marks.
<IP address> Either an IP address in dotted decimal form or a host name
string that occurs in the local hosts file or on an Internet
domain name server.
The RFCNAMES.LST file can be modified manually, or by using MPTS and
selecting the Names list option when you Edit configuring IBM OS/2
NetBIOS OVER TCP/IP. This is shown in Figure 181 on page 240 and Figure
182 on page 240.
Figure 181. [OS/2 Warp] NetBIOS over TCP/IP configuration options
Figure 182. [OS/2 Warp] Modifying RFCNAMES.LST through MPTS
240
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The advantage of using a names file is the reduction of name query
broadcasts. Before the computer broadcasts to the network for NetBIOS
name to IP address resolution, it searches this local file. The search is done
by checking each names file entry up to the length of the entry. Due to the
search technique used, the order of entries in the names file is important.
If an address in the file is given as a host name string, it is resolved to an IP
address by looking it up in the local hosts file or by querying the Internet
domain name server. The host name string must, therefore, be in the same
form here as it is in the hosts file or on the domain name server.
For an OS/2 Warp Server environment, the requesters need a names file with
the computer name, IP address and domain name IP address of each server
they contact. The servers need the computer name and IP address for each
requester that contacts them. The computer name and domain names are
found in the IBMLAN.INI files on these machines.
Broadcast file routing extension
This routing extension is implemented through the use of a broadcast file that
contains the IP addresses of other subnets. A TCP/IP broadcast is limited to
one subnet unless other subnet addresses are included in the broadcast file.
The broadcast file, named RFCBCST.LST, is located in the \IBMCOM
directory. It contains a list of host names (such as merlot), host IP addresses
(such as 192.168.6.10) or subnet broadcast addresses (such as
192.168.6.255). The file is read once at startup and each valid address is
added to the set of destination addresses for broadcast packets. Any host or
network accessible by these addresses becomes part of the local NetBIOS
broadcast domain; that is, they receive all name lookup and registration
packets, as well as all broadcast and group NetBIOS datagrams. These
addresses can be on other networks and accessed through the appropriate
router or bridge by the IBM TCP/IP program. The remote node is treated as if
it were a node on the local network.
Each line of the broadcast file is either an IP address in dotted decimal
notation or a host name string that occurs in either the local hosts file used by
the IBM TCP/IP program or on an Internet domain name server. The host
name string must be in the same form here as it is in the hosts file or on the
domain name server.
Again, the broadcast list may be edited manually or through MPTS.
Chapter 5. Integrating file and print services
241
Figure 183. [OS/2 Warp] Modifying RFCBCST.LST through MPTS
For OS/2 Warp Server and OS/2 Warp File and Print Clients, if the server and
requester are on different networks, the broadcast file must be updated. At
the clients, the broadcast file must include the IP address or TCP/IP host
name for each server that this requester accesses using TCPBEUI. At the
servers, the broadcast file must include the directed broadcast address for
each subnet containing LAN requesters.
The following must apply when using a broadcast file:
• IP addresses, subnet broadcast addresses, and host name strings are
allowed.
• A maximum of 32 entries is allowed.
Note
IBM may provide a Service Pack in 2000 that enables you to work with
more than 32 entries.
• Host name strings should match those in the hosts file and domain name
server.
• Routers should be checked to ensure that they forward directed
broadcasts.
RFCADDR.EXE program
When you make changes to the names/broadcast files while TCPBEUI is
active, you can reinitialize TCPBEUI with the new file by using the RFCADDR
command. The cache will be updated with the new names.
242
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Domain Name Server (DNS) routing extension
This routing extension uses the TCP/IP Domain Name Server (DNS), to store
NetBIOS names and IP address pairs, so that a TCPBEUI client can obtain
the IP address of another workstation from it. This extension requires that the
workstations already have an IP name and address registered on the DNS.
The network administrator then enters the NetBIOS name and the IP address
of every TCPBEUI workstation into the DNS.
To enable the use of the Domain Name Server routing extension, you need to
set the DOMAINSCOPE parameter in the PROTOCOL.INI file equal to your
TCP/IP domain name. For example:
DOMAINSCOPE=austin.cooking.net
If the names file search fails, the domain scope string, if specified, is
appended to the encoded NetBIOS name. The new host name string is then
used to resolve the NetBIOS name to an IP address by looking up the name
in the local TCP/IP hosts file or by querying the DNS. If the DNS knows the
name, it sends back the IP address to the NetBIOS over TCP/IP program.
The advantage of this implementation is that you have a central repository of
all NetBIOS names. However, you have to update your DNS with NetBIOS
names that are in an encoded format. An encoded format is necessary
because NetBIOS names can be 16 bytes of any bit pattern, but the domain
name server only accepts a limited character set. The TCPBEUI program
encodes them into the 32-byte, reversible, half-ASCII format that is specified
in RFCs 1001/1002.
You can use the MAPNAME.EXE utility to encode NetBIOS names.
MAPNAME is supplied in the \APPLETS\MPTSAPLT.ZIP file on MPTS
diskette 5.
MAPNAME syntax
MAPNAME <input> /FDBLxx
Where:
<input>
Either filename.LST or filename.RFC NetBIOS name; maximum of
16 characters for a NetBIOS name, and 32 characters for a RFC
name
Name can be any char.+ \xDD in any order, surrounded by "[]",
D is a hex digit (\xDD = one char.)
Chapter 5. Integrating file and print services
243
F
Indicates input from file
(default is name)
D
Indicates the mapping direction
Use r to indicate NetBIOS to RFC
Use n to indicate RFC to NetBIOS
B
Append blanks up to the 16 character
(default is a null character)
Lxx
Indicates the last character
(xx can be any hex value)
***
If F, then the name can be the combination of any character that
can be put in the file by the editor
Once the names have been encoded, you can store them in the DNS (or
DDNS) so that they point back to the original host name. Remember that
NetBIOS names are registered multiple times (with a different suffix) when a
computer starts. So, each instance of name+suffix must be encoded and
registered on the DNS.
Note
The NetBIOS and DOMAIN names must be encoded in upper-case when
using MAPNAME.
Here’s an extract from the DNS database before the encoded NetBIOS
names are added.
MERLOT
86400
IN
IN
A
192.168.6.10
HINFO Domain Cont.
;
BRIE
86400
IN
IN
A
192.168.7.12
HINFO Workstation
;
Figure 184. [Warp Server] Sample DNS database
Using the Warp Server MERLOT as an example, its name is registered three
times with the 0x00, 0x03 and 0x20 suffixes. In addition, being on the domain
ARMONK, it also registers the domain name with a 0x00 suffix.
Using MAPNAME for each instance of name+suffix, we encode the names
into the following:
MAPNAME MERLOT /rbl00 gives the encoding ENEFFCEMEPFECACACACACACACACACAAA
244
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
MAPNAME MERLOT /rbl03 gives the encoding ENEFFCEMEPFECACACACACACACACACAAD
MAPNAME MERLOT /rbl20 gives the encoding ENEFFCEMEPFECACACACACACACACACACA
MAPNAME ARMONK /rbl00 gives the encoding EBFCENEPEOELCACACACACACACACACAAA
Workstations, on the other hand, would normally register their names three
times. They would register the NetBIOS name with the 0x00 and the 0x03
suffixes. Usually, they would also register the domain name with the 0x00
suffix as well, but seeing as this encoded name is already registered on the
DNS, it cannot be registered again.
Using MAPNAME for the workstation BRIE, we encode the following:
MAPNAME BRIE /rbl00 gives the encoding ECFCEJEFCACACACACACACACACACACAAA
MAPNAME BRIE /rbl03 gives the encoding ECFCEJEFCACACACACACACACACACACAAD
The encoded names can now be added to the DNS database and pointed
back to the TCP/IP host name (Using CNAME), where the IP address is
specified. An extract from the new DNS database now looks like this:
MERLOT
86400
IN
IN
A
HINFO
192.168.6.10
Domain Cont.
;
ENEFFCEMEPFECACACACACACACACACAAA
86400
IN
IN
CNAME
HINFO
MERLOT
0x00 suffix
;
ENEFFCEMEPFECACACACACACACACACAAD
86400
IN
IN
CNAME
HINFO
MERLOT
0x03 suffix
;
ENEFFCEMEPFECACACACACACACACACACA
86400
IN
IN
CNAME
HINFO
MERLOT
0x20 suffix
;
EBFCENEPEOELCACACACACACACACACAAA
86400
IN
IN
CNAME
HINFO
MERLOT
Domain 0x00
;
BRIE
86400
IN
IN
A
HINFO
192.168.7.12
Workstation
;
ECFCEJEFCACACACACACACACACACACAAA
86400
IN
IN
CNAME
HINFO
BRIE
0x00 suffix
;
ECFCEJEFCACACACACACACACACACACAAD
86400
IN
IN
CNAME
HINFO
BRIE
0x03 suffix
Figure 185. [Warp Server for e-business] DNS database with encoded NetBIOS names
Chapter 5. Integrating file and print services
245
5.3 New way of using DNS with non-RFC-encoded name
The previous section informed you how to create the RFC-encoded names for
server names and domain names. It is a huge effort for TCP/IP administrators
to set up all NetBIOS names as static entries in the DNS files. Basically, the
NetBIOS name server, as defined in the RFC 1001/1002, is the ideal solution
for applications written to the NetBIOS over TCP/IP interface.
However, there is a new way of NetBIOS name resolution on the horizon:
• Non-RFC-encoded names in DNS; the node status request frame was
introduced as a communication vehicle
This new technique is introduced through Microsoft’s Windows NT Service
Pack 3 and can be implemented on Windows NT Workstation clients, as well
as Windows NT Server. OS/2 Warp Server for e-business is able to support
this technique and as far as OS/2 Warp Server Version 4 is concerned, there
is a fix (APAR IC20646) for the server, but OS/2 Warp clients do not yet
support this new technique.
Figure 186 illustrates the setup and interaction of the Windows NT server and
the Windows NT Workstation client.
DNS
Server
SERVER-A
DOMAIN-A
1. Query DNS forDOMAIN-A
2. Get IP Address of DOMAIN-A
A 9.53.4.5
CNAME SERVER-A
OS/2 Warp Server
for e-business
Windows NT
Client
3. Query NetBIOS
names with Node
Status Request
frame
4. Return all
NetBIOS names,
including 0x1B, Ox1C,
0x1D, and so forth
Windows NT
Server
DNS Only Setup
5. Logon
Figure 186. Using DNS server with node status request
In the DNS domain file, the Windows NT server computername, also known
as server name, is appended as a pure ASCII name with an A record that
points to the server’s static IP address. In addition, the domain name is
246
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
included with an alias, CNAME, to the server’s name. The entries to the DNS
domain file would look like the following two lines:
SERVER-A
DOMAIN-A
IN
IN
A
CNAME
9.53.4.5
SERVER-A.
The logon sequence would use the following steps:
1. The Windows NT Workstation client queries the DNS server for hostname
DOMAIN-A.
2. The DNS server returns the information that DOMAIN-A is the alias to
SERVER-A. The Windows NT Workstation client receives the IP address
of the Windows NT Server.
3. Since the Windows NT Workstation client is not sure whether or not the
target IP address is really a NetBIOS node that has certain NetBIOS
names registered, it sends a RFC frame to the received IP address (which
is the Windows NT Server). This frame is called Node Status request and
passes along a resource record name of * (asterisk).
4. Windows NT Server responds to the Node Status request by sending a
complete list of NetBIOS names registered in the node.
5. Now the Windows NT Workstation client selects the target IP address that
has a correct server name and domain name, then it starts the logon
sequence through a SMB session.
This technique might be called DNS-Only or Use of non-RFC-encoded names
NetBIOS names resolution, but it actually is a combination of DNS name
resolution and the Node Status Request frame defined in the RFC
1001/1002. Neither Windows NT Server nor Windows NT Workstation client
use WINS in this case. The TCP/IP properties are set up as shown in Figure
187 on page 248 and Figure 188 on page 249.
Chapter 5. Integrating file and print services
247
Figure 187. [Windows NT] WINS address no use of WINS, but use DNS
Windows NT will warn you when you don’t provide the WINS server
parameter at all. However, you can ignore this message and have the
workstation configured to work with DNS only.
248
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 188. [Windows NT] TCP/IP Property - DNS settings
Figure 189 on page 250 shows the two IP frames, a packet 49 sent from a
Windows NT workstation with service pack 3 carrying a Node Status request
frame with a questioned name *, and a packet 50 frame that was sent back
from a Windows NT Server showing all registered NetBIOS names and status
on the node.
The questioned name *(asterisk), followed by 15 bytes of null (0x00)
characters, is actually translated into the RFC-encoded name that results in a
character string of CKAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA in the
trace data.
Chapter 5. Integrating file and print services
249
Packet 49: IP,
9.3.1.138
-> 9.3.1.188
Source IP: 9.3.1.138
, Destination IP: 9.3.1.188
Version: 04,
IP header length: 05 (32 bit words)
Service type: 0: Precedence: 0, Delay: Norm, Throug: Norm, Reliab: Norm
PROTOCOL: [17] UDP
UDP,
[137] -> [137]
Source port: [137] ,
Destination port: [137]
TCPIP NETBIOS (name service)
Operation: [0] NAME QUERY REQUEST, Flags: Unicast,
Transaction ID: 0803
Header Counts: QD: 1, AN: 0, NS: 0, AR: 0
Request Name: *
Request type: NODE STATUS REQUEST
RAW PACKET LISTING:
0000 18 40 00 04 AC EC 35 D0 80 00 83 F4 6B 13 06 40
[email protected]¬ì5D.ô[email protected]
0010 40 31 40 10 AA AA 03 00 00 00 08 00 45 00 00 4E
@[email protected]ªª......E..N
0020 81 29 00 00 20 11 04 2B 09 03 01 8A 09 03 01 BC
).. ..+......¿
0030 00 89 00 89 00 3A A2 E0 08 03 00 00 00 01 00 00
...:¢à........
0040 00 00 00 00 20 43 4B 41 41 41 41 41 41 41 41 41
.... CKAAAAAAAAA
0050 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41
AAAAAAAAAAAAAAAA
0060 41 41 41 41 41 00 00 21 00 01
AAAAA..!..
--------------------------------------------------------------Packet 50: IP,
9.3.1.188
-> 9.3.1.138
Source IP: 9.3.1.188
, Destination IP: 9.3.1.138
Version: 04,
IP header length: 05 (32 bit words)
Service type: 0: Precedence: 0, Delay: Norm, Throug: Norm, Reliab: Norm
PROTOCOL: [17] UDP
UDP,
[137] -> [137]
Source port: [137] ,
Destination port: [137]
TCPIP NETBIOS (name service)
Operation: [0] NAME QUERY RESPONSE, Flags: Unicast,
Transaction ID: 0803
Header Counts: QD: 0, AN: 1, NS: 0, AR: 0
Resource Record Name: *
Request type: NODE STATUS RESOURCE RECORD
Data
0000 54 53 4F 4E 54 30 30 20 20 20 20 20 20 20 00 44
TSONT00
.D
0010 00 49 54 53 4F 41 55 53 4E 54 20 20 20 20 20 20
.ITSOAUSNT
0020 00 C4 00 49 54 53 4F 4E 54 30 30 20 20 20 20 20
.Ä.ITSONT00
0030 20 20 20 44 00 49 54 53 4F 41 55 53 4E 54 20 20
D.ITSOAUSNT
0040 20 20 20 20 1C C4 00 49 54 53 4F 41 55 53 4E 54
.Ä.ITSOAUSNT
0050 20 20 20 20 20 20 1B 44 00 49 54 53 4F 41 55 53
.D.ITSOAUS
0060 4E 54 20 20 20 20 20 20 1E C4 00 49 54 53 4F 4E
NT
.Ä.ITSON
0070 54 30 30 20 20 20 20 20 20 20 03 44 00 49 54 53
T00
.D.ITS
0080 4F 41 55 53 4E 54 20 20 20 20 20 20 1D 44 00 01
OAUSNT
.D..
0090 02 5F 5F 4D 53 42 52 4F 57 53 45 5F 5F 02 01 C4
.__MSBROWSE__..Ä
00A0 00 49 4E 65 74 7E 53 65 72 76 69 63 65 73 20 20
.INet~Services
00B0 1C C4 00 49 53 7E 49 54 53 4F 4E 54 30 30 00 00
.Ä.IS~ITSONT00..
00C0 00 00 00 44 00 49 54 53 4F 4E 54 30 30 20 20 20
...D.ITSONT00
00D0 20 20 20 20 01 44 00 41 44 4D 49 4E 49 53 54 52
.D.ADMINISTR
00E0 41 54 4F 52 20 20 03 44 00 00 04 AC EC 35 D0 00
ATOR .D...“5D.
00F0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
................
0100 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
................
0110 00 00 00 00 00 00 00 2B C3 80 B0 3E C3 80 14 2C
.......+ð>Ã.,
0120 C3 80 0F 00 01 00 00 00 00 00 00 00 00 00 00 00
Ã..............
0130 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
................
0140 00 00 00 00 00 00 00 00 00
.........
Figure 189. [Windows NT] Packet trace - Node status request
250
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
When we used an NT Workstation with Service Pack 5 to log on to an NT
Server also with Service Pack 5, we found an interesting surprise. The query
to the DNS Server was using WINS packets even though there was no WINS
server or client enabled anywhere on the network. Figure 190 shows one of
the frames we captured using Sniffer Basic. Note how there is no use made of
a * character as there was with Service Pack 3 for Windows NT.
- - - - - - - - - - - - - - - - - - - Frame 9 - - - - - - - - - - - - - Frame Source Address
Dest. Address
Size Rel. Time
Delta Time
9 PIZZA
[192.168.8.255]
102 000:00:06.276 1.025.598
Abs. Time
Summary
10/19/1999 12:34:52 PM WINS: C ID=32774 OP=QUERY NAME=NTNP2300<1C>
: DLC AC=18, FC=40, FS=00
RI: Broadcast Fwrd
LLC: C D=AA S=AA UI
SNAP: Ethernet Type=0800 (IP)
IP: D=[192.168.8.255] S=[192.168.8.60] LEN=58 ID=2304
UDP: D=137 S=137 LEN=58
WINS: ----- WINS Name Service header ----WINS:
WINS: ID = 32774
WINS: Flags = 01
WINS: 0... .... = Command
WINS: .000 0...
= Query
WINS: .... ..0. = Not truncated
WINS: .... ...1 = Recursion desired
WINS: Flags = 1X
WINS: ...1 .... = Non Verified data is acceptable
WINS: Question count = 1, Answer count = 0
WINS: Authority count = 0, Additional record count = 0
WINS:
WINS: Question section:
WINS:
Name = NTNP2300<1C> <Domain Controller>
WINS:
Type = NetBIOS name service (WINS) (NetBIOS name,32)
WINS:
Class = Internet (IN,1)
WINS:
ADDR HEX
ASCII
0000: 18 40 ff ff ff ff ff ff 80 20 35 21 c8 60 82 70 | [email protected] 5!.‘.p
0010: aa aa 03 00 00 00 08 00 45 00 00 4e 09 00 00 00 | ........E..N....
0020: 80 11 9f 13 c0 a8 08 3c c0 a8 08 ff 00 89 00 89 | .......<........
0030: 00 3a 85 62 80 06 01 10 00 01 00 00 00 00 00 00 | .:.b............
0040: 20 45 4f 46 45 45 4f 46 41 44 43 44 44 44 41 44 | EOFEEOFADCDDDAD
0050: 41 43 41 43 41 43 41 43 41 43 41 43 41 43 41 42 | ACACACACACACACAB
0060: 4d 00 00 20 00 01
| M.. ..
Figure 190. Frame 9 captured with Sniffer Basic
As shown in Figure 191, the NBTSTAT command shows the formatted output of
the Node Status Request frame.
Chapter 5. Integrating file and print services
251
C:\>NBTSTAT -A 192.168.8.60
NetBIOS Remote Machine Name Table
Name
Type
Status
--------------------------------------------PIZZA
<00> UNIQUE
Registered
NTNP2300
<00> GROUP
Registered
PIZZA
<03> UNIQUE
Registered
PIZZA
<20> UNIQUE
Registered
NTNP2300
<1E> GROUP
Registered
ADMINISTRATOR <03> UNIQUE
Registered
MAC Address = 00-20-35-21-C8-60
Figure 191. [Windows NT] NBTSTAT output
The CIFS (Common Internet File System) protocol as defined in the IETF
draft, is supposed to extend the Internet to support the SMB (Server
Message Block) file sharing protocol over the Internet. The idea of CIFS is
the same as illustrated before:
• Non-RFC-encoded names in DNS
• Node Status request frame as communications vehicle
To make CIFS work properly, the name of *SMBSERVER followed by the blanks
must not be used as a NetBIOS name by servers or workstations in a
network. Using CIFS, clients do not have knowledge of server names or
domain names. All SMB servers should respond to the *SMBSERVER name but
not through a network broadcast. Clients are able to check whether or not a
particular TCP/IP host is an SMB server.
The CIFS Internet draft is located on the Internet at the following Web site:
http://search.ietf.org/internet-drafts/draft-leach-cifs-v1-spec-02.txt
However, due inactivity this draft has expired, as can be seen when
consulting the following Web site: http://www.cifs.com/spec.html
This is still in informational status. There is a common belief that CIFS would
lack security and that authentication logic is not strong enough on the public
Internet. In addition to CIFS, remote administration and the network
neighborhood browser protocol are discussed.
252
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
5.4 NetBIOS Name Servers (NBNS)
It would be much simpler if, instead of having to manually maintain
configuration files on each computer, we could ask a NBNS to resolve the
NetBIOS name to an IP address for us.
We will look at two types of NBNS. The first is WINS from Microsoft, the
second is Shadow IPserver from Network TeleSystems.
5.4.1 Microsoft WINS
WINS is Microsoft’s implementation of NBNS. WINS only supports Microsoft's
proprietary clients with its implementation of native NetBIOS and NetBIOS
over TCP/IP
5.4.1.1 How WINS works
Each Microsoft client needs to be configured with the IP address of a primary
WINS server, and optionally with the IP address of a secondary WINS server.
Whenever a client (configured to use TCPBEUI and WINS) starts, it will
attempt to register its NetBIOS name and IP address with the primary WINS
server. The registration occurs when services or applications are started (for
example, Workstation or Messenger), and is sent directly to the primary
WINS server. If the name is not already registered to another client, the
server responds with a message detailing the NetBIOS name that has been
registered, and the Name Time to Live (TTL).
If after attempting three times to register its name with the primary server and
failing, the client will attempt to register its name using the secondary server.
If the secondary server also fails to respond, the client will revert to
broadcasting in order to register its name.
The name registrations are made on a temporary basis, and the client is
responsible for maintaining the lease of the registered name.
At one-eighths of the TTL, the client will attempt to refresh its name
registration with the primary WINS server. If the client does not receive a
response from the server, it will continue to attempt to refresh the registration
every two minutes until half the TTL has expired. At this point, it will repeat
the procedure, but this time using the secondary WINS server.
With WINS enabled, the client acts as an H-Node client for name registration.
For resolution, it is H-Node with a few modifications. The sequence used by a
WINS client for name resolution is:
Chapter 5. Integrating file and print services
253
• Check to see if it is the local machine name.
• Check the cache. (Any resolved name is placed in a cache for 10 minutes)
• Try to use the primary WINS server. (Use the secondary server if the
primary does not answer after three attempts.)
• Try a name query broadcast.
• Check the LMHOSTS file (if the computer is configured to use LMHOSTS).
• Try the HOSTS file.
• Try the DNS.
5.4.1.2 System requirements for WINS
Table 10 details the minimum system requirements to implement a WINS
server.
Table 10. WINS hardware and software requirements
Type
Details
Hardware requirements
Processor
80486-33 (Pentium recommended)
Hard drive
125 MB
Memory
16 MB (32 MB recommended)
Network adapter
Any network adapter supported by Windows NT
Video
VGA or above
Software requirements
Windows NT
NT Server 4.0 with Service Pack 5
IP configuration
IP address
Subnet mask
Default router
(these can be allocated by DHCP, but statically assigned
parameters are recommended)
5.4.1.3 Installing a WINS server
(This procedure assumes that Windows NT has previously been installed,
and that the TCP/IP connectivity on the network is working)
1. Click on the Start menu. Select Settings, then select Control Panel.
2. Double-click on the Network icon.
254
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
3. Now select the Services tab.
4. Select the Add button.
5. Select Windows Internet Name Service from the list.
6. Select the OK button.
7. The WINS service will be installed.
8. If you had a Service Pack installed on the server before you installed the
WINS Server, then you must re-install the Service Pack.
9. You may optionally add static entries for non-WINS clients.
Non-WINS clients can be statically added so that WINS clients can resolve
their name without having to maintain a LMHOSTS file.
10.You may also optionally set up a WINS proxy agent.
A WINS proxy agent extends the functions of a WINS server to non-WINS
clients. When the proxy agent sees a name registration request, it
forwards the request to a WINS server. The server verifies the name only,
but does not register it. When the proxy agent sees a query to resolve a
name, it forwards the query to the WINS server. The server responds to
the agent, who, in turn, returns the details to the client.
Note
The IP addresses manually configured for primary and secondary WINS
servers take precedence over those supplied as DHCP options.
5.4.1.4 WINS limitation
According to RFC 1001/1002, a NetBIOS name server should support all
group names. WINS, however, only keeps a list of IP addresses for group
names ending in 0x1C. Warp Server domains, however, are registered with a
0x00 suffix, and, as such, are not stored by a WINS server. Therefore, when
an IBM client requests an IP address from WINS server, a broadcast address
(an IP address of all ones) is returned. This makes it difficult for IBM clients to
communicate across a routed network when using WINS as the NBNS.
5.4.2 Network TeleSystems Shadow IPserver
Shadow IPserver is a software system for managing name and address
assignments and host configuration information within a TCP/IP network.
IPserver includes a robust set of network services offered through
standards-based protocol interactions:
• DHCP (Dynamic Host Configuration Protocol) service
Chapter 5. Integrating file and print services
255
• DNS (Domain Name System) service
• NBNS (NetBIOS Name Server) and NBDD (NetBIOS Datagram
Distributor) services
The Shadow IPserver NBNS service is an integrated NetBIOS name server
and datagram distributor. The IPserver NBNS service fully implements RFCs
1001 and 1002. Additionally, Shadow IPserver’s NBNS service provides
name resolution services similar to those provided by Microsoft WINS.
We used the NTS Shadow IPserver in our test networks as a fully integrated
DHCP, DNS and NBNS/WINS server. For details on the installation of
IPserver and the configuration of its services, please refer to Part 6.4,
“Dynamic IP with Network TeleSystems Shadow IPserver” on page 293.
Note
To use the datagram distributor, both the NBNS server and TCP/IP client
stack must support datagram distribution, such as:
• OS/2 Warp Server with latest MPTS ServicePak
• OS/2 Warp 4
• DOS LAN Services (shipped with OS/2 Warp Server)
The Shadow IPserver NBNS server supports both IBM and Microsoft
NetBIOS networks; the same NBNS server can be used with OS/2 Warp,
DOS/Windows workstations with DOS LAN Services installed, Windows 95,
and Windows NT workstations.
The idea of the NBNS is that instead of broadcasting the NetBIOS names to
the whole IP subnet, the names to be registered are sent to a NBNS server.
The NBNS server then checks if the name is already in use, and if not, the
name is registered. All the future queries to that name will be resolved with
one single query to the NBNS.
To take full advantage of the your NetBIOS name server, you can use the IBM
Neighborhood Browser Enabler, which is included as a service within Warp
Server for e-business. The browser service allows the OS/2 Domain
Controller to register the Microsoft network dependent NetBIOS names and
acts as the domain master browser machine. See Part 5.6, “IBM
Neighborhood Browser Enabler for Warp Server” on page 271 for information
on Neighborhood Browser Enabler.
256
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
5.5 P-Node, H-Node and M-Node clients
After installing a NBNS, all the servers and clients running NetBIOS over
TCP/IP need to be configured to use the NBNS server. Notice that servers
and clients normally register their NetBIOS names when the requester
services are started.
Clients can have the IP address of the NBNS manually coded, or they can
receive the address as options with their DHCP configuration.
5.5.1 Manually configuring clients to use an NBNS
The following information demonstrates how to manually configure a client to
make use of a NBNS.
5.5.1.1 DOS, Windows 3.1, and Windows for Workgroups
To have DOS and Windows clients running DLS communication with an
NBNS (rather than using the NBUTIL utility), the \NET\PROTOCOL.INI file
must be modified with the addition of statements to the [NTS$NTSTS] (or
[TCPPRO2]) sections.
Note
These statements cannot be added or configured through the DLS
installation program.
Use a text editor such as EDIT or NOTEPAD to modify the PROTOCOL.INI to
add the following options
NODETYPE=H-Node
NBNSAddr=[NBNS_ipaddress]
NBDGAddr=[NBDD_ipaddress]
Where:
[NBNS_ipaddress]
Is the IP Address of the NBNS
[NBDD_ipaddress]
Is the IP Address of the Datagram Distributor
Figure 192 on page 258 shows an extract from a PROTOCOL.INI file, with the
statements to support communications with a NBNS highlighted.
Chapter 5. Integrating file and print services
257
[NTS$NTST2]
DNSAddr=
GatewayAddr=
NetSubNetMask=
IPAddr=
TCPHeartBeats=Standard
DHCPClientID=BURGER
BootPFlag=DHCP
NODETYPE=H-Node
NBNSAddr=[192.168.6.5]
NBDGAddr=[192.168.6.5]
DriverName=ntsts$
VCs=16
VCReceiveLarge=6
VCSends=6
RcvWindow=2920
UseMemory=UMB
BINDINGS=IBM$GENIBMTOK
LANABASE=0
Token-Ring
Figure 192. [DOS LAN Services] PROTOCOL.INI extract showing NBNS setup
5.5.1.2 Windows 95 and Windows 98
In Windows 95 and Windows 98, the NBNS server IP address must be
defined as a WINS server. The operating system takes care of the proper
node-type selection. Set the WINS address the following way:
1. Go to the Control Panel and double-click the Network icon.
2. Select the TCP/IP protocol from the list and click on the Properties button.
3. Select the WINS Configuration tab, and select Enable WINS
Resolution. See Figure 193 on page 259.
258
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 193. [Windows 98] Manually configuring a NBNS
4. Enter the NBNS server IP address as the primary WINS server.
5. Optionally enter the IP address for a backup NBNS as the secondary
WINS server.
6. Save the changes and reboot the machine to make the new settings
effective.
5.5.1.3 Windows NT
In Windows NT, the NBNS server IP address must be defined as a WINS
server. The operating system takes care of the proper node-type selection.
Set the WINS address the following way:
1. Go to the Control Panel and double-click the Network icon.
2. Select the Protocols tab.
3. Select the TCP/IP protocol from the list and click on the Properties button.
4. Select the WINS Address tab. See Figure 194 on page 260.
Chapter 5. Integrating file and print services
259
Figure 194. [Windows NT] Manually configuring a NBNS
5. Enter the NBNS server IP address as the primary WINS server.
6. Optionally enter the IP address for a backup NBNS as the secondary
WINS server.
7. Save the changes and reboot the machine to make the new settings
effective.
5.5.1.4 OS/2 Warp
In order to have OS/2 Warp 4 clients communicate with a NetBIOS Name
Server, you need to make changes to the default TCPBEUI settings. Using
MPTS (Multi-Protocol and Transport Services) follow the steps below:
1. From System Setup (Warp Center pull-down menu) select MPTS. You can
also open an OS/2 Window, and type MPTS at the command prompt.
2. Select Configure at the Multi-Protocol Transport Services window.
3. Select the radio button for LAN adapters and protocols (which is usually
set by default) and select Configure.
4. In the Adapter and Protocol window, make sure the current configuration
list contains the IBM OS/2 NetBIOS over TCP/IP protocol. If IBM OS/2
NetBIOS is listed, the number for IBM OS/2 NetBIOS over TCP/IP must be
260
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
changed to 2 by selecting the Change number... button. Double-click on
IBM OS/2 NetBIOS over TCP/IP.
5. At the following NetBIOS over TCP/IP window, select the radio button for
Driver parameters, then click on Configure.
6. At the Parameters for IBM OS/2 NetBIOS OVER TCP/IP window, make the
following changes (Figure 195).
• Set the Node Type parameter to H-Node or P-Node, depending on the
need. In most situations it is recommended to use H-Node. If P-Node is
used, registrations cannot be done without NBNS server. If H-Node is
used, the registrations are broadcast if the NBNS server is not
available.
• Set the NetBIOS Name Server address parameter to the NBNS server
IP address.
• Set the NetBIOS Datagram Distribution server parameter to the NBNS
sever IP address.
• Optionally set the backup NBNS and NBDD server IP addresses.
Figure 195. [OS/2 Warp Version 4] Manually configuring a NBNS
Chapter 5. Integrating file and print services
261
7. Select OK, Close, OK, Close and Exit. Allow MPTS to update your
CONFIG.SYS, and click on Exit again. You will be informed that the
config.sys was updated. Click on Ok. Click on Exit for the final time to
close MPTS.
8. To make the protocol changes active, you must reboot your workstation.
5.5.2 Dynamically configuring clients to use a NBNS
A DHCP server is able to supply NBNS server information to DHCP clients
using standard options. There are three options available to inform clients of
the NBNS servers.
Option
Description
044 NBNS
The addresses of the primary and (optionally) the backup
NBNS servers
045 NBDD
The addresses of the primary and (optionally) the backup
NetBIOS Datagram Distributor servers
046 Node type
0x1 = B-Node, 0x2 = P-Node, 0x4 = M-Node, 0x8 = H-Node
Note that pure WINS clients do not make use of a Datagram Distributor
server, so option 045 may not be required in your environment.
5.5.2.1 DOS, Windows 3.11 and Windows for Workgroups
With DLS installed, all that is required is to ensure that PROTOCOL.INI does
not contain the NODETYPE, NBNSAddr and NBDGAddr parameters.
The computer will then use options 44, 45, and 46 when it receives them.
5.5.2.2 Windows 95 and Windows 98
In Windows 95, the computer needs to be configured to use a returned NBNS
IP address:
1. Go to the Control Panel and double-click the Network icon.
2. Select the TCP/IP protocol from the list and click the Properties button.
3. Select the IP Address tab and choose to Obtain an IP address
automatically.
4. Click on the WINS Configuration tab and select Use DHCP for WINS
Resolution.
5. Click on OK to save the changes. Reboot the machine to make the new
settings active.
262
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
5.5.2.3 Windows NT
With Windows NT, ensure that no WINS servers have been manually
configured. Any manually configured NBNS servers have priority over NBNS
IP addresses returned as DHCP options.
5.5.2.4 OS/2 Warp
At this point in time, OS/2 Warp Connect and OS/2 Warp 4 do not directly
support options 44, 45, and 46. So, we looked into a workaround for this
problem and came up with the following solution.
The solution uses two REXX scripts: SETNBNS.CMD and GETNBNS.CMD.
The SETNBNS script is run by the DHCP client when it receives the NBNS
options. To do this, the \MPTN\ETC\DHCPCD.CFG file requires the following
entries:
option 44 exec "setnbns.cmd 44 %s"
option 45 exec "setnbns.cmd 45 %s"
option 46 exec "setnbns.cmd 46 %s"
When the options are received, SETNBNS creates a REXX queue, and places
the parameters into the queue. Figure 196 on page 264, Figure 197 on page
265, and Figure 198 on page 266 display SETNBNS.CMD.
Chapter 5. Integrating file and print services
263
/****************************************************************************
* SETNBNS.CMD
* by Peter Degotardi
* Revised April 21, 1998
* SETNBNS
* Return values: 0=successful, 1=unsuccessfull
* Accept DHCP options 44, 45 and 46 and put them in a queue where
* GETNBNS.CMD can read them.
****************************************************************************/
return_code = 1
/* Assume everything will fail */
/*** add RexxUtil functions ***/
call RxFuncAdd 'SysLoadFuncs', 'RexxUtil', 'SysLoadFuncs'
call SysLoadFuncs
/* get option number and data from the command line */
parse arg option_tag option_data
if ((option_tag = '') | (option_data = '')) then
do
return_code = 1
/* wrong number of arguments */
exit return_code
end
/*** strip leading and trailing spaces ***/
option_tag = strip(option_tag)
option_data = strip(option_data)
Select
when (option_tag = 44) then
/* NBNS address */
do
if (check_IPaddr(option_data)) then
return_code = queue_it(option_tag||','||option_data)
end
Figure 196. SETNBNS.CMD (Part 1 of 3)
264
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
when (option_tag = 45) then
do
if (check_IPaddr(option_data)) then
/* NBDD address */
return_code = queue_it(option_tag||','||option_data)
end
when (option_tag = 46) then
do
call lineout '\setnbns.log', 'option_data="'option_data'"'
select
when option_data = 8 then data = 'H-Node'
when option_data = 4 then data = 'M-Node'
when option_data = 2 then data = 'P-Node'
otherwise data = 'B-Node'
end
return_code = queue_it(option_tag||','||data)
end
otherwise
nop
/* unsupported_option */
end
exit return_code
/****************************************************************************
* check_IPaddr
* Check a dotted decimal IP address for valid format.
*
* Return 0 if there are not 4 octets or the data is out of range or
* there are embedded blanks in the address, else return 1.
****************************************************************************/
check_IPaddr: Procedure
parse arg addr
addr = strip(addr)
parse var addr octet.1 '.' octet.2 '.' octet.3 '.' octet.4
do pos = 1 to 4
if (verify(octet.pos, '0123456789', N) \= 0) then
return 0
/* not an unsigned number */
if ((octet.pos < 0) | (octet.pos > 255)) then
return 0
end
return 1
/****************************************************************************
* hex2ascii_string (option_data)
*
* Take a hex string in the form 'hex”nn nn nn ... ", verify it's a valid
* hex string, and convert it to an ASCII character string.
*
* Returns the string if the data is valid, a null string if not.
****************************************************************************/
hex2ascii_string: Procedure
parse arg 'hex' '"' data '"'
if (\ DATATYPE(data, X)) then return ""
data = strip(data, 'B', '09'x)
return X2C(data)
Figure 197. SETNBNS.CMD (Part 2 of 3)
Chapter 5. Integrating file and print services
265
/****************************************************************************
* queue_it
* Put the data in a queue where it can be read by GETNBNS.
* Tries to create the queue. If a different name is returned, the queue
* must already exist, so delete the 'wrong' queue, set the queue to the
* 'correct' one, and put the data there.
****************************************************************************/
queue_it: Procedure
parse arg data
qname = 'NBNSINFO'
/* The queue we're using */
new_queue = rxqueue("create", qname)
if new_queue \= qname then
rc = rxqueue("delete", new_queue)
rc = rxqueue("set", qname)
queue data
return 0
Figure 198. SETNBNS.CMD (Part 3 of 3)
GETNBNS.CMD is the second script and is called from STARTUP.CMD with
GETNBNS /GO. (We use the /GO parameter so that users do not accidently run
the script.) It reads the queue created by SETNBNS and will rewrite
PROTOCOL.INI if new parameters have been received. The script will wait a
maximum of 60 seconds for new parameters, and will give up if none are
received.
If you have rewritten PROTOCOL.INI, a reboot will be required to use the new
options. Rather than automatically rebooting the computer, a message is
displayed informing the user to do so.
Figure 199 on page 267, Figure 200 on page 268, Figure 201 on page 269,
and Figure 202 on page 270 display GETNBNS.CMD.
266
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/****************************************************************************
* GETNBNS.CMD
* by Peter Degotardi
* Revised April 21, 1998
* GETNBNS /GO
* Return values: 0=successful, 1=no TcpBEUI, 2=Can't read PROTOCOL.INI
*
-1=timeout waiting for TcpBEUI options
*
99=run without specifying the /GO parameter
*
* Read DHCP options 44, 45 and 46 from the queue created by SETNBNS.CMD
*****************************************************************************/
/* Mandatory flag so user's don't accidently run this */
arg run_flag
if run_flag \= '/GO' then exit 99
return_code = 1
is_data = 1
are_changes = 0
nbns_queue = 'NBNSINFO'
prot.0 = 0
/* Assume everything will fail */
/* There might be data */
/* Have changes been made to PROTOCOL.INI */
/* The queue input queue holding NBNS data */
/* Stem for holding PROTOCOL.INI in memory */
/*** add RexxUtil functions ***/
call RxFuncAdd 'SysLoadFuncs', 'RexxUtil', 'SysLoadFuncs'
call SysLoadFuncs
/* Where's PROTOCOL.INI? */
boot_drive = filespec('drive', value('comspec', ,'os2environment'))
prot_ini = boot_drive || '\IBMCOM\PROTOCOL.INI'
/* Does this machine have TcpBEUI ? */
if tcpbeui_chk() = 0 then
do
call delete_queue
exit 1
end
rc = rxqueue("set", nbns_queue)
if queued() = 0 then
do t = 60 to 10 by -10
is_data = 0
/* Nope, no data */
say 'Waiting 't' seconds for new TcpBEUI options.'
call syssleep(10)
if queued() \= 0 then
/* We've seen some data, so */
do
/* get out of here */
is_data = 1
leave
end
end
if is_data then
do
if read_prot() = 0 then
do forever
if queued() = 0 then leave
parse pull item
parse value item with option','data
data = strip(data)
Figure 199. GETNBNS.CMD (Part 1 of 4)
Chapter 5. Integrating file and print services
267
select
when option = 44 then
do
call set_nbns data
end
when option = 45 then
do
call set_nbdd data
end
when option = 46 then
do
call set_node data
end
otherwise
nop
end
call SysSleep(5)
end
else
exit 2
/* Give time for more options to arrive */
/* Couldn't read PROTOCOL.INI */
end
else
do
say 'No new TcpBEUI parameters received.'
say 'Giving up.'
exit -1
end
if are_changes then
do
rc = write_prot()
say 'Changes have been made to TCPBEUI options in PROTOCOL.INI.'
say 'Please reboot to activate them.'
say ''
say 'Press Enter to continue...'
pull entkey
end
exit 0
/****************************************************************************
* tcpbeui_chk
* Check to see if tcpbeui is configured on this machine. It's not worth
* continuing if it isn't
****************************************************************************/
tcpbeui_chk:
call SysFileSearch prot_ini, '[TCPBEUI_NIF]', has_tcpbeui
return has_tcpbeui.0
/****************************************************************************
* delete_queue
* Delete the queue created by SETNBNS
****************************************************************************/
delete_queue:
rc = rxqueue("delete", nbns_queue)
return
Figure 200. GETNBNS.CMD (Part 2 of 4)
268
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/****************************************************************************
* read_prot
* Read PROTOCOL.INI into memory
****************************************************************************/
read_prot:
rc = stream(prot_ini, 'c', 'open read')
if rc = 'READY:' then
do
p = prot.0
do while lines(prot_ini) \= 0
p = p + 1
prot.0 = p
prot.p = linein(prot_ini)
end
rc = stream(prot_ini, 'c', 'close')
return 0
end
else
return 1
/****************************************************************************
* write_prot
* Write the changed PROTOCOL.INI to disk
****************************************************************************/
write_prot:
prot_bak = prot_ini|| '.OLD'
call SysFileDelete(prot_bak)
/* Not checking for ReadOnly file */
'@COPY 'prot_ini' 'prot_bak' >NUL'
/* Not checking for OK copy */
call SysFileDelete(prot_ini)
/* Not checking for OK delete */
rc = stream(prot_ini, 'c', 'open write')
if rc = 'READY:' then
do
do p = 1 to prot.0
call lineout prot_ini, prot.p
end
rc = stream(prot_ini, 'c', 'close')
say 'Writing 'prot.0' lines to 'prot_ini
return 0
end
else
return 1
/****************************************************************************
* set_node
* Set the Node Type in PROTOCOL.INI if different.
****************************************************************************/
set_node:
parse arg nodetype
do p = 1 to prot.0
if left(strip(translate(prot.p)), 8) = 'NODETYPE' then
do
parse value prot.p with .'=' current_type
parse value current_type with '"'current_type'"'
if translate(current_type) \= translate(nodetype) then
do
are_changes = 1
prot.p = '
NODETYPE = "' || nodetype || '"'
Figure 201. GETNBNS.CMD (Part 3 of 4)
Chapter 5. Integrating file and print services
269
say 'NODETYPE was 'current_type', replacing with 'nodetype
end
end
end
return
/****************************************************************************
* set_nbns
* Set the NBNSADDR in PROTOCOL.INI if different.
****************************************************************************/
set_nbns:
parse arg nbns_addr
do p = 1 to prot.0
if left(strip(translate(prot.p)), 8) = 'NBNSADDR' then
do
parse value prot.p with .'=' current_addr
parse value current_addr with '"'current_addr'"'
if current_addr \= nbns_addr then
do
are_changes = 1
prot.p = '
NBNSADDR = "' || nbns_addr || '"'
say 'NBNSADDR was 'current_addr', replacing with 'nbns_addr
end
end
end
return
/****************************************************************************
* set_nbdd
* Set the NBDDADDR in PROTOCOL.INI if different.
****************************************************************************/
set_nbdd:
parse arg nbdd_addr
do p = 1 to prot.0
if left(strip(translate(prot.p)), 8) = 'NBDDADDR' then
do
parse value prot.p with .'=' current_addr
parse value current_addr with '"'current_addr'"'
if current_addr \= nbdd_addr then
do
are_changes = 1
prot.p = '
NBDDADDR = "' || nbdd_addr || '"'
say 'NBNSADDR was 'current_addr', replacing with 'nbdd_addr
end
end
end
return
Figure 202. GETNBNS.CMD (Part 4 of 4)
Technical caveats
Be aware that the above scripts are not perfect, but are aimed to give you an
idea as to how to process these options. We, therefore, draw your attention to
some of the issues.
• The SETNBNS.CMD script will run every time a lease is renewed. As such, the
REXX queue NBNSINFO will be recreated again and data will be queued.
270
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
With a short lease time on a machine that is rarely restarted, you could
end up with a queue containing a large quantity of data.
• Unfortunately, sometimes when the OS/2 Workstation is booting up, the
queue is flushed before the OS/2 desktop starts up. So, GETNBNS.CMD does
not have any parameters to update in the PROTOCOL.INI. However, when the
TCP/IP lease expires and is renewed while the machine is up and working,
SETNBNS.CMD will be re-run.Then GETNBNS.CMD /GO should be run again and
the PROTOCOL.INI will be updated if there is anything in the queue.
• Options 44 and 45 can contain multiple IP addresses and SETNBNS and
GETNBNS cannot cope with this.
• Normally, the exec entries in DHCPCD.CFG for a REXX script would be:
option 44 exec "$setnbns.cmd 44 %s"
The $ parameter is used to force a 30 second wait so that REXX can
initialize. We found that by including the $ parameter, a 30 second wait
was introduced between each instance of SETNBNS running for each option,
that is, a total of 90 seconds was required if all three options were
processed. With this delay, GETNBNS only saw the first option queued.
Without the $ parameter, SETNBNS and GETNBNS worked correctly.
• Error handling could be greatly improved. As you can see from the
comments in GETNBNS, the backup/deletion/re-writing of PROTOCOL.INI
could be improved.
• GETNBNS can be run from the Startup folder, but be aware that the user may
not see the message that a restart is required.
• If the user chooses not to reboot, they may not be able to log onto the
LAN.
5.6 IBM Neighborhood Browser Enabler for Warp Server
IBM Neighborhood Browser Enabler for OS/2 Warp Server, now included in
OS/2 Warp Server for e-business, enables greater interoperability between
Microsoft clients and OS/2 Warp Servers.
5.6.1 Overview
The Neighborhood Browser Enabler enables your OS/2 Warp Server to
function as a master browser for Windows 95, Windows 98 and Windows NT
clients. If you are using Windows 95, Windows 98, or Windows NT clients in
an OS/2 Warp Server domain, and have not started the Neighborhood
Browser Enabler, then you have to set up one master browser per segment to
be able to browse OS/2 Warp Server domain resources from Windows 95,
Chapter 5. Integrating file and print services
271
Windows 98, and Windows NT. The master browser workstation has to be up
and running all the time.
We recommend enabling the Neighborhood Browser Enabler on the domain
controller machine. Start it on any OS/2 Warp Server for e-business machine,
one master browser at a time. Starting Neighborhood Browser Enabler on
more than one server reduces the network traffic between the master browser
and the clients. All the other servers running the Neighborhood Browser
Enabler service are backup browsers. For performance considerations, it is
recommended to enable Neighborhood Browser Enablers for every 20
servers in the domain. Depending on the installation, this figure may vary.
5.6.2 Installing neighborhood browser enabler
The following sections describe the installation and configuration of the IBM
Neighborhood Browser Enabler on OS/2 Warp Server for e-business and
OS/2 Warp Server Version 4.
5.6.2.1 OS/2 Warp Server for e-business
The Neighborhood Browser Enabler is installed with a LAN server during the
installation process of OS/2 Warp Server for e-business.
The browser service does not start automatically when the server service is
started.
The IBMLAN.INI file contains a new section:
[browser]
maintainserverlist = YES
quiet= No
It also adds a new service called BROWSER in the SRVSERVICES line, and
it extends the path to a new executable in the [services] section, namely:
browser = services\browser.exe
5.6.2.2 For OS/2 Warp Server Version 4
To install the Neighborhood Browser Enabler on OS/2 Warp Server Version 4,
follow these steps:
1. Download the Neighborhood Browser Enabler as a ZIP file from the IBM
Software Choice at:
http://service.software.ibm.com/asd-bin/doc/index.htm
(This is a free download.)
272
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
2. From an OS/2 Window on your Warp Server, create a temporary directory,
and copy the downloaded ZIP file into it.
3. Unzip the package to the directory by using PKUNZIP2:
PKUNZIP2 BRINST.ZIP
4. Read the README.1ST file.
5. If the Server is running, stop it by entering the command:
NET STOP SERVER
6. Enter the command:
BRINST
This will install new files into their appropriate directories:
\IBMLAN\SERVICES\BROWSER.EXE
\IBMLAN\NETLIB\BROWSE32.DLL
\IBMLAN\NETLIB\BROWSE16.DLL
Depending on the service level, selected base LAN Server modules are
backed up, then newer versions are installed:
\IBMLAN\SERVICES\NETSERVR.EXE
\IBMLAN\SERVICES\NETSVINI.EXE
\IBMLAN\NETPROG\NETWKSTA.200
\IBM386FS\HPFS386.IFS
\MUGLIB\DLL\NETAPI32.DLL
Note: The file \MUGLIB\DLL\NETAPI32.DLL is replaced only on OS/2 Warp
Server SMP IP08500.
7. A reboot may be necessary depending on LAN Server's service level.
The Neighborhood Browser Enabler installation program copies the browser
files to the appropriate directories after backing up the existing IBMLAN.INI
file to IBMLAN.IBR, backing up some LAN server components, and finally
modifying the IBMLAN.INI file so that the browser service starts automatically
when the server service is started.
The updated IBMLAN.INI file contains a new section:
[browser]
maintainserverlist = auto
It also adds a new service called BROWSER in the SRVSERVICES line, and
it extends the path to a new executable in the [services] section, namely:
browser = services\browser.exe
Chapter 5. Integrating file and print services
273
5.6.3 Starting and stopping the Neighborhood Browser Enabler
You can start the Neighborhood Browser Enabler service at any time by
issuing the command:
NET START BROWSER
Or, stop the service by using the command:
NET STOP BROWSER
You may also stop and start the Neighborhood Browser Enabler service
through the LAN Server Administration GUI (select LAN Server
Administration => Local Workstation => Services => Browser => Start or
Stop).
When the Neighborhood Browser Enabler is running on the OS/2 Warp
Server for e-business domain controller machine, the domain controller also
recognizes and answers to NetBIOS name types used by Microsoft networks
but not by IBM networks. These types include domain controller, domain
master browser, master browser, and browser service elections. In cases
where Microsoft Windows clients are attached to the network, we recommend
starting the Neighborhood Browser Enabler and using the appropriate IBM
Networks Client for Windows.
5.7 AIX Fast Connect
AIX Fast Connect 2.1 allows an RS/6000 to function as a file or print server to
Windows 95/98/NT and OS/2 clients. In this section we will cover some
scenarios where we tested Fast Connect with some DHCP settings. For more
information on the implementation and function of Fast Connect, see the
redbook titled AIX and Windows NT: Solutions for Interoperability,
SG24-5102.
5.7.1 Fast Connect on static server with dynamic client
In our lab, we set up an RS/6000 Model 43P with AIX 4.3.3 as a Fast Connect
and DHCP server. We set up a Windows NT workstation to obtain DHCP
lease from our RS/6000 and then define a mapped network drive to the
RS/6000. Then, using the command ipconfig /release, we had our Windows
NT workstation release its DHCP lease. After that we executed the command
stopsrc -s dhcpsd on the RS/6000 to stop DHCP server services, and then
started DHCP services on an NTS Shadow IP server. Finally, we had our
Windows NT workstation obtain a new lease from the Shadow IP server and,
without fail, we still had connectivity to our drive space defined on the Fast
Connect server. In addition, we were able to define new network drive space.
274
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
For client concerns, DHCP and Fast Connect services do not interfere with
one another and there are no DHCP concerns with a static Fast Connect
server and dynamic Windows clients.
5.7.2 Fast Connect on a dynamic server with dynamic client
When we ran Fast Connect server on an AIX DHCP client we ran into
difficulties. We had our AIX system get a DHCP lease and IP address of
192.168.6.201 and then started the Fast Connect server and defined drive
space on that server from a Windows NT client. Then, we released our DHCP
lease on the AIX Fast Connect server, got a new lease with a new IP address
of 192.168.6.50, and then tried to view the Fast Connect server in the Map
Network Drive menu on the Windows NT client. While trying to view the Fast
Connect server on the new IP lease, the Windows NT client hung and we had
to end the task. However, stopping and starting the Fast Connect server, and
therefore having it listen on a valid IP address, solved the problem and clients
were now able to connect to the Fast Connect server without fail.
In the event that you are going to run Fast Connect on an AIX DHCP client,
we recommend that the DHCP client reserve its IP address on the DHCP
server, run the Fast Connect server on static IP, or be prepared to restart Fast
Connect services in the event the lease expires on your AIX DHCP client.
Chapter 5. Integrating file and print services
275
276
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 6. Expanding your network
Planning and installing a simple DHCP and DDNS network is not a difficult
task. However, when we need more DHCP servers or more (D)DNS servers,
there are many complex design issues. We discuss multiple subnets, multiple
DHCP servers, address pool management, DHCP relay agents, combination
of a static DNS and DDNS, a mix of NT DNS and OS/2 Warp Server DDNS,
and AIX DDNS. If your network is growing, you will want to read this chapter.
6.1 Multiple subnets
There are several reasons why you would divide your TCP/IP network with
subnets. Assume a small company, HIC Corp., had been using one floor of
the office building downtown, and now the company expands one more floor.
If the newly added floor is next to the original floor, they might just extend the
LAN so that a bridge or repeater is enough to extend the TCP/IP network.
However, if the LAN utilization is already high, they may want to use a router
to physically divide the two LANs, thus giving two subnets, one for each
physical floor of the building where they have a group of machines on the
network. Furthermore, in scenarios where you are setting up large numbers
of workstations in a vast complex, you may want to subdivide the workstation
groups, either by floor or by area, by their subnet, to make the IP topology
resemble the physical layout of your LAN more closely.
6.1.1 Static and dynamic routing between subnets
Routing is the process of choosing a network path on which to transfer
packets. A router is a device that is (usually) connected to two or more
networks and is responsible for transferring data from one network to another.
In order to perform routing, a router has to maintain a map of what subnets
are on the network and where it should send data so that it can reach its
destination. There are two ways of maintaining this list:
• The network administrator could manually configure the routing table on
each router. This is said to be static routing. Static routing is easy to
administer if your network is fairly small, but can become incredibly
complex as your network continues to grow.
• The routers could talk amongst themselves to determine what subnets are
on the network and how to get to them. Two common protocols used by
routers are Routing Information Protocol (RIP) and Open Shortest Path
First (OSPF).
© Copyright IBM Corp. 1998, 2000
277
Both OS/2 Warp and Windows NT include the functionality to act as routers.
They can be configured as static or dynamic routers (or a combination of
both).
Note
OS/2 Warp and Windows NT only support RIP. If you have routers that use
OSPF (or other routing protocols), OS/2 Warp and Windows NT will not be
able to exchange routing information with the other routers.
In showing how to configure OS/2 Warp and Windows NT as routers, we’ll
assume the sample network shown in Figure 203.
Figure 203. Sample network for routing
The routing tables required for this sample network are shown in Table 11.
Table 11. Routing tables required for sample network
278
Router 1 Routing table
Router 2 Routing table
Data for Network
Send to
Data for network
Send to
192.168.7.0
192.168.7.1
192.168.7.0
192.168.6.1
192.168.6.0
192.168.6.1
192.168.6.0
192.168.6.4
192.168.9.0
192.168.6.4
192.168.9.0
192.168.9.1
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
6.1.2 OS/2 Warp as a router
To enable OS/2 Warp 4 with TCP/IP Version 4.1 to act as a router, it should
be connected to two or more networks and have statically assigned IP
addresses for each network adapter.
6.1.2.1 Enabling routing
To enable routing under OS/2 Warp 4, do the following:
• Using the TCP/IP configuration panels, switch to the Routing tab, and
check the IP Forwarding checkbox, as shown in Figure 204.
Figure 204. [OS/2 Warp] Enabling IP forwarding
• Click on OK, and keep exiting until you have exited from the TCP/IP
configuration program.
If you examine the \MPTN\BIN\SETUP.CMD file, you will see that the line:
IPGATE OFF
has been changed to:
IPGATE ON
• Shut down and restart.
Chapter 6. Expanding your network
279
6.1.2.2 Static routing
To configure static routes under OS/2 Warp, do the following:
• Using the TCP/IP configuration panels, switch to the Routing tab, and click
on the Add button. The Route Entry panel is displayed.
As shown in Figure 205, we are configuring Router 1 so that it knows
where to send packets destined for the 192.168.9.0 subnet.
Figure 205. [OS/2 Warp] Adding a route
• Click on OK, and keep exiting until you have exited from the TCP/IP
configuration program.
If you now examine the \MPTN\BIN\SETUP.CMD file, you will see that any
routes that you have added from the Route Entry panel have been
inserted as route commands. For example, the entry for router 2 will be:
route add -net 191.168.7.0 192.168.6.1 netmask 255.255.255.0 -hopcount 1
• Shut down and restart.
ROUTE Syntax
ROUTE [ -nqv] [ add | delete] [ -net | -host] dest gateway [-netmask mask]
[-nqv] [ change] [ -net | -host] dest [gateway] [-lock | -lockrest]
[-rtt | -rttvar | -sendpipe | -recvpipe | -mtu | -hopcount | -expire |
-ssthresh] value
[-nqv] [ get] [-net | -host] dest
[-nqv] [ monitor]
[-nqv] [ flush]
[-nqv] [ add | delete | change | get] [default] gateway
280
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Where:
-n
Bypass translating IP addresses to symbolic hostnames.
-q
Suppress all output.
-v
Verbose.
add
To add a route.
delete
To delete a route.
change
To change aspects of a route.
get
To look up the route for a destination.
monitor
To continuously report changes to routes.
flush
To remove all routes.
-net
The following dest is a network address.
-host
The following dest is a hostname or address (default).
-netmask
The following is the mask of the route.
dest
The IP address or hostname of the destination.
gateway
The IP address or hostname of the next-hop router.
mask
The subnet mask.
value
A value.
default
0.0.0.0, for all destinations not defined by any other routes.
6.1.2.3 Dynamic routing
For Dynamic routing, TCP/IP makes use of the ROUTED command. The Routed
command enables RIP, so that the computer can interact with other routers.
• Using the TCP/IP configuration panels, switch to the Autostart tab.
• From the Autostart Services window, select routed, as shown in Figure
206 on page 282.
Chapter 6. Expanding your network
281
Figure 206. [OS/2 Warp enabling ROUTED
• Select the Autostart Service checkbox.
• Select either the Detached or Foreground Session radio button.
• If you selected Foreground Session, optionally select the Minimized
checkbox.
• Optionally, set startup parameters. Refer to the online documentation for
parameters that may be used with routed.
• Click on OK, and keep exiting until you have exited from the TCP/IP
configuration program. If you now examine the
\TCPIP\BIN\TCPSTART.CMD file, you will see that a start routed
command has been added.
• Shut down and restart.
6.1.3 Windows NT as a router
As with OS/2 Warp, Windows NT should be connected to two or more
networks with statically assigned IP addresses for each adapter before it can
act as a router.
282
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
6.1.3.1 Enabling routing
From the Control Panel, open the Network icon; then open the Properties for
the TCP/IP protocol.
Select the Routing tab, and check the Enable IP Forwarding checkbox as
shown in Figure 207.
Figure 207. [Windows NT] enabling IP forwarding
Click on OK and exit from the network configuration.
6.1.3.2 Static routing
To add static routes to the routing table, you use the route command.
• For example, to permanently add the route for the 192.168.7.0 subnet to
router 2, issue the command:
route -p add 192.168.7.0 mask 255.255.255.0 192.168.6.1
Chapter 6. Expanding your network
283
ROUTE syntax
ROUTE [ -p] add network mask mask gateway metric metric
delete network gateway
change network gateway
print
-f
Where:
-p
Makes persistent changes; that is, changes are stored to the
registry
add
Adds a route
delete
Deletes a route
change
Changes a route
network
The destination subnet
mask
The subnet mask
gateway
Where to send packets destined for the subnet network
metric
The cost of using this route
-f
Flush the routing table
6.1.3.3 Dynamic routing
To perform dynamic routing, the RIP for Internet Protocol Service must be
installed. This service comes as part of Windows NT Server.
1. Open the Network icon in the Control Panel, and on the Services tab,
select Add. A list of available services will be displayed.
2. Scroll down the list, and select RIP for Internet Protocol (as shown in
Figure 208 on page 285); then click OK.
284
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 208. [Windows NT] installing the RIP for Internet Protocol service
• Keep exiting until you have exited from the network configuration.
• Shut down and restart.
6.2 Crossing routers
Once your TCP/IP network begins to grow, you will, for various reasons, split
the network into multiple subnets that are connected through routers.
Routers generally filter out all broadcast traffic between subnets. With this in
mind, you should remember that a workstation that is being configured
through DHCP is dependent on the DHCPDISCOVER and subsequent
broadcasts being seen by a DHCP server.
So, how do you make your DHCP clients work on one subnet when your
DHCP server is on another subnet and your broadcast traffic is being filtered?
The simplest option is to configure your router to pass on all broadcast traffic.
This will work, but will introduce further problems as your network continues
to grow. Therefore, is not recommended.
A better method would be to have some type of device on the subnet remote
from the DHCP server listening for DHCP broadcasts, and when it sees one,
it forwards it directly to the DHCP server.
RFC1542 Clarifications and Extensions for the Bootstrap Protocol documents
the methodology for implementing such a device. Although it was originally
written with BOOTP in mind, it is valid for DHCP. A device that implements
this type of forwarding is known as a DHCP Relay Agent (or a BOOTP Relay
Chapter 6. Expanding your network
285
Agent). Figure 209 shows the initial setup sequence of a DHCP client when
the DHCP server is remote, but a Relay Agent is available.
Figure 209. DHCP initialization through a DHCP relay agent
You may have RFC 1542-compliant routers; if so, the Relay Agent can be run
internally on the router. If your routers are not compliant, or if you do not want
to add any additional processing load to the routers, there are software
solutions available in Windows NT and OS/2 Warp to make standard
computers act as DHCP Relay Agents.
Important
Before you consider running a DHCP Relay Agent on a machine that is
itself a DHCP client, remember that it must be able to contact a DHCP
server to get an IP address. It won't be able to do this because the Relay
Agent is not running. Therefore, the Relay Agent should be run on a
machine with a statically assigned IP address.
6.2.1 Windows NT as a DHCP Relay Agent
The DHCP Relay Agent is bundled with the Windows NT Server, but is not
installed by default.
286
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
To install, open the Network icon in the Control Panel, and on the Services
tab, select Add. A list of available services will be displayed with the DHCP
Relay Agent at the top of the list. Select it, and click on OK.
Figure 210. [Windows NT] List of available network services
The Relay Agent will be installed and added to the list of installed services.
Click on Close. An error panel will be displayed stating that the DHCP Relay
Agent cannot be used until a DHCP server IP address is configured. Click on
Yes to add a DHCP server.
Figure 211. [Windows NT] error panel shown after relay agent install
The TCP/IP properties panel is displayed. Select the DHCP Relay tab.
Chapter 6. Expanding your network
287
Figure 212. Windows NT] DHCP relay tab
Click on the Add button. A panel will be displayed that will allow you to enter
an address for a DHCP server. You can add multiple DHCP servers.
Figure 213. Windows NT] adding a DHCP server
When you have finished adding the DHCP server addresses, click on the OK
button to exit. As usual, you will have to reboot.
Figure 214. Windows NT] reboot prompt
288
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
6.2.2 OS/2 Warp as a DHCP Relay Agent
Starting with TCP/IP Version 4.1 for OS/2 Warp, IBM has bundled a DHCP
Relay Agent. Although information on how to set up the DHCP Relay Agent is
not provided in the online documentation, you can obtain setup instructions
as APAR IC20987 from the IBM Support Center or on the Web (by searching
the APAR database) at:
http://ps.software.ibm.com.
For your convenience, we have also provided those setup instructions in this
section.
The Relay Agent consists of one executable, \TCPIP\BIN\DHCPRD.EXE, and
a required configuration file. By default, the configuration file,
\MPTN\ETC\DHCPRD.CFG, is used. A sample configuration file may be
found in \TCPIP\SAMPLES\ETC, as shown in Figure 215.
numLogFiles
logFileSize
logFileName
logItem
logItem
logItem
logItem
logItem
logItem
logItem
logItem
logItem
server
#
4
100
dhcprd.log
SYSERR
OBJERR
PROTERR
WARNING
EVENT
ACTION
INFO
ACNTING
TRACE
9.67.116.156
Figure 215. [OS/2 Warp sample DHCPRD.CFG file
The configuration file must be manually edited using a text editor such as E or
TEDIT. The options that can be contained within the file are as follows:
#
Comment
numLogFiles
The number of log files desired.
logFileSize
The size of log files in K bytes.
logFileName
The name of the most recent log file.
logItem
One item to be logged. The loggable items are:
Chapter 6. Expanding your network
289
SYSERR
Log system errors.
OBJERR
Log object errors between objects in the process.
PROTERR
Log protocol errors between client and server.
WARNING
Log warnings deserving attention from the user.
EVENT
Log events that occur to the process.
ACTION
Log actions taken by the process.
INFO
Log information that might be useful.
ACNTING
Log of who was served when.
TRACE
Log code flow for debugging.
server
IP address of a single DHCP server. The server option may be
specified multiple times so that DHCP requests are forwarded
to multiple servers.
DHCPRD usage
DHCPRD [ -?] | [-b] | [ [-v | -q] [ConfigFile] ]
Where:
-?
Displays help message.
-b
Displays the program banner.
-q
Executes in quiet mode.
-v
Executes in verbose mode.
ConfigFile
Relay Agent configuration file. This file defaults to the
following rules:
Name
dhcprd.cfg
Location
Searches the current working directory. If it is not there, the
%ETC% directory (C:\MPTN\ETC) is searched.
Once the configuration file has been created, the Relay Agent can be started
in several ways:
• The DHCPRD command may be added to \STARTUP.CMD.
• The DHCPRD command may be added to \TCPIP\BIN\TCPSTART.CMD.
• An icon can be added to the Startup Folder.
290
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
6.2.3 DHCP relay considerations
6.2.3.1 The broadcast flag
DHCP includes a method whereby devices unable to receive a packet with a
specific IP address can ask the server or relay agent to use the broadcast IP
address for replies (a flag set by the client in the requests). The definition of
DHCP states that implementations should honor this flag, but it doesn't say
they must. Some TCP/IP implementations used this flag, which meant that
Relay Agents and servers had to implement it.
A number of BOOTP Relay Agent implementations (for example, in routers)
handled DHCP correctly except for this feature.
6.2.3.2 Virtual LANs
Some virtual LAN (VLAN) schemes use a data packet's IP address to decide
which virtual LAN a device is on. Depending on the capabilities of your DHCP
server, this can represent a significant challenge to your implementation.
DHCP servers and Relay Agents use their knowledge of what LAN the device
is physically connected to in order to select the subnet number for the
device's new IP address. Whereas, with switches, the subnet number sent by
the device is used to decide which virtual LAN to put the device on.
6.2.3.3 Multiple subnets on a single wire
Networks are sometimes configured so that one network segment has
multiple subnets running on it. The DHCP server allocates addresses from a
pool that is defined for each subnet or network.
In this type of environment, unless the DHCP Relay Agent is able to indicate
which subnet a device requesting an address belongs to (by indicating to the
server its own IP address for that subnet), or unless the server itself can be
configured to cope with this environment, then you will probably not be able to
implement DHCP.
6.3 IP masquerading/network address translation
One way to combat an IP address shortage is to avoid assigning addresses in
the first place. Yet without an IP address, no system can communicate on a
TCP/IP network.
IP masquerading allows you to share one IP address among several different
systems while still providing reliable, full-function TCP/IP connections to
every host. This technique can be used to provide Internet service to more
than one computer using a single dial-up connection from an Internet Service
Chapter 6. Expanding your network
291
Provider (ISP), for example. Or, you may wish to link branch office systems to
your company’s network without assigning IP addresses from your company’s
main address pool.
In fact, you still must assign an IP address to every host that needs to
communicate on the TCP/IP network, but the IP addresses may be from a
private network pool such as the Class A network 10. A masquerading
gateway accepts TCP/IP connection requests from these private hosts and
translates them into requests originating from the IP address assigned to the
masquerading gateway itself. Then, responses are automatically routed back
to the correct host on the private network. As a result, a single IP address on
one network can effectively serve several different hosts without requiring
additional IP address assignments on that same network.
InJoy from F/X Communications can provide IP masquerading services on
your network, perhaps in combination with Virtual Private Network (VPN)
capabilities. (VPN can securely link intranets at different sites using the public
Internet.)
If you are making intermittent connections to the Internet, perhaps Network
Address Translation (NAT) is for you. For example, assume you have a small
network at home, or you have a small office with light use of the Internet.
There are a number of different products that allow you to share a modem
over a regular telephone line or an ISDN line.
Sometimes NAT is referred to as IP masquerading, although NAT is a super
set of IP Masquerading (the IP address does not actually change as it passes
from one LAN/WAN to another through the masquerade software). Originally,
NAT was suggested as a short-term solution to the problem of IP address
depletion. In order to be assured of any-to-any communication in the Internet,
all IP addresses have to be officially assigned by the Internet Assigned
Numbers Authority (IANA). This is becoming increasingly difficult to achieve,
because the number of available address ranges is now severely limited.
Also, many organizations have, in the past, used locally assigned IP
addresses, not expecting to require Internet connectivity.
The idea of NAT is based on the fact that only a small part of the hosts in a
private network are communicating outside of that network. So, if we can
devise a technique to assign official addresses to hosts that are used only
when they need to communicate outside the private network, then only a
small number of official addresses are required.
This is what NAT does; it takes the IP address of an outgoing packet and
dynamically translates it to an official address. For incoming packets, it
292
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
translates the official address to an internal address. We now can use NAT for
a solution for networks that have private address ranges or illegal addresses
and want to communicate with hosts on the Internet.
In fact, by implementing a firewall, we have already circumvented part of the
problem. Clients that communicate on the Internet by using a proxy or socks
server do not expose their addresses to the Internet, so their addresses do
not have to be translated anyway. However, when we do not want to use a
proxy or sock server, or when proxy and socks are not possible, we can use
NAT. For example, proxy and socks servers cannot be used for UDP
connections on some firewall products.
6.3.1 Translation mechanism
For each outgoing IP packet, the source address is checked by the NAT
configuration rules. If a rule matches the source address, the address is
translated to an official address from the address pool. The predefined
address pool contains the addresses that NAT may use for translation. For
each incoming packet the destination address is checked to see if it is used
by NAT. When this is true the address is translated to the original unofficial
address.
Many different products implement Network Address Translation or IP
Masquerading. Often true NAT is available with firewall products. One
inexpensive product for OS/2 for IP Masquerading is the INJOY dialer. It also
provides Dial On Demand. It can be used with modems over regular phone
lines or ISDN.
For more information on IP masquerading with InJoy, please visit the Web
site:
http://www.fx.dk
6.4 Dynamic IP with Network TeleSystems Shadow IPserver
As your network continues to grow, you may find it beneficial to use dedicated
hardware or software to provide the functions of DHCP, DDNS, and NetBIOS
name resolution. Network TeleSystems, (see http://www.nts.com) offers
several solutions in this area. We employed their Shadow IPserver Version
3.04o in our test networks to provide fully integrated DHCP, DNS, and
NetBIOS name resolution services. We also used Network TeleSystems
IPmanager software for Windows to aid in the configuration and management
of the server.
Chapter 6. Expanding your network
293
6.4.1 Installing Shadow IPserver
Table 12 shows the system requirements to implement Shadow IPserver on a
standalone PC. NTS also offers a version of IPserver that runs as a service
on Windows NT, as well as a fault tolerant, rack-mounted hardware platform.
Table 12. Shadow IPServer hardware and software requirements
Type
Details
Hardware requirements
Processor
80486-100 or faster
Hard Drive
150 MB IDE/EIDE or larger (SCSI is not supported)
Memory
16 MB
Video
VGA
Network Adapter
One of the following network adapters:
3Com 3C509 ISA Ethernet
Eagle Technology’s Novell NE2000plus3 ISA Ethernet
Intel EtherExpress Pro/100+ (PILA 8460) PIC adapter
(82558 chip)
IBM Auto 16/4 Token-Ring ISA
Note: You can use the IBM Token Ring Turbo 16/4 if you use
the LANAID software provided by IBM to configure the card
in 16/4 ISA16 mode.
Software requirements
Operating System
IBM or MS DOS 5.0 or later (IBM PC DOS 2000 is Y2K
Ready)
Note: Although IPserver uses DOS for installation and
startup, it interacts with the PC hardware components at a
very low level to achieve the best performance. The
config.sys and autoexec.bat files should contain only those
items recommended by NTS. Other programs could
interfere with the performance and reliability of the server.
IP Configuration
IP address
Subnet mask
Default router
The configuration used for our test environment was:
• Shadow IPserver Version 3.04o
• IBM PC-DOS 2000
• IBM PC 365 (Pentium Pro 180 Mhz)
• 64 MB memory
• An IBM Turbo 16/4 Token-Ring Adapter, configured manually:
294
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• 16 Mb/sec data rate
• ROM address C800
• IO port 0A20
6.4.2 Installation procedure
The IPserver installation program asks you for configuration information and
creates the NTS-SRVR.CFG file based on your input. To install IPserver,
follow the steps below:
1. Ensure the machine you’ve selected to install Shadow IPserver on meets
the following requirements:
1. Drive C: is your primary disk partition.
2. Drive C: must be formatted and installed with DOS.
3. IPserver requires HIMEM.SYS to be loaded via your CONFIG.SYS file.
Advanced memory management programs like EMM386.EXE or
RAMBOOST.EXE should not be used. If these files are being referred
to within the CONFIG.SYS or AUTOEXEC.BAT files, the statements
must be removed.
4. Check the PC's time and date to verify that they are correct. Changes
made after your server is running might affect its configuration. Refer to
6.4.3, “IPserver configuration overview” on page 298 for details.
2. Place the Shadow Installation diskette into the PC's diskette drive.
3. Start the installation program by issuing the following command:
<drive letter>:\INSTALL
where <drive letter> is the drive that contains the installation diskette.
4. When prompted, enter the required information such as the IP address
you’ve selected for the server, the subnet mask and the default router
address.
5. When prompted, select the appropriate network adapter and adapter
settings. You need to know the port address and memory range your
network adapter is using, as the installation program will ask for this
information. These parameters can also be changed later, and are located
in the NTS-SRVR.CFG file in the C:\SHADOW directory.
At a minimum, the server needs to be configured with the correct
parameters for the network adapter card, IP address and subnet mask,
and default gateway. With these parameters configured you have the
option of using NTS’s IPmanager software, or a Java enabled browser to
perform further configuration. The IPmanager program runs on Windows
Chapter 6. Expanding your network
295
and uses a graphical user interface to configure and monitor multiple
servers.
During the installation of Shadow IPserver, you have the option of installing a
sample configuration. This is can be helpful if you’re learning how to configure
the product for the first time. For our installation, we did not install the sample
configuration.
At the completion of the installation program, you can review or change the
information you entered by editing the NTS-SRVR.CFG file in the
C:\SHADOW directory. Figure 216 on page 297 shows the NTS-SRVR.CFG
we file used in our test environment.
296
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
#OutAHere
NoDiags
No-Promiscuous
Replace-Client-ID
PING
History
BOOTP
DHCP
WINS-Mode
IPAddr 192.168.6.5
NetSubnetMask 255.255.255.0
BroadCastAddr All-1s
GateWayAddr 192.168.6.1
ArpTimeout 2
IPTimeToLive 4
Restart 0
AutoSave-Time 0
SyncTTL 10
BackupLimit 10
BackupTime 10
CoServerTTL 100000000
TTL 10800
ReleaseTTL 10
DNSBackupLimit 0
DNSBackupTime 0
NegCacheTTL 60
PosCacheTTL 100000000
DHCPBackupLimit 0
DHCPBackupTime 0
TCP-Idle-Time 100
TCP-Retransmit 100
Ping-Positive-Timeout 299
Ping-Negative-Timeout 299
Ping-Retries 1
Ping-Timeout 200
DNSCoServer-Retries 2
DNSCoServer-Timeout 500
DNSNameServer-Retries 2
DNSNameServer-Timeout 500
Time-Offset 36
SNMP-Get-Community public
SNMP-Set-Community public
NIC IBMTR
Connector TPI
Window 0C800
IO-Port 0a20
Figure 216. [Shadow IPserver] NTS-SRVR.CFG configuration file
The settings we chose to change for our test environment were:
• The OUTAHERE parameter was remarked. This prevents Shadow from
exiting if you accidentally hit the Esc key. To exit Shadow press
Ctrl-Right Shift-Esc.
• The TTL (Time To Live) setting was set to 10800 seconds (three hours) to
give optimal support for Windows 95 clients.
Chapter 6. Expanding your network
297
6. Check that the command NTS-SRVR was added to the AUTOEXEC.BAT file
to start IPserver. If you are using Shadow’s IPmanager software to
configure your server you can start the server and proceed from the GUI
screens. If you prefer to edit the textual configuration files, continue with
steps 7 and 8.
7. Optionally, modify the NTS-SRVR.NBN file to add entries to the NBNS
database on the server at startup.
8. To manually configure DHCP and DDNS, modify the NTS-SRVR.DHC and
NTS-SRVR.DNS files as required.
6.4.3 IPserver configuration overview
You can configure IPserver by manually editing its textual configuration files,
by using the management software (IPmanager), or both. When you use
IPmanager, your configuration entries are stored in several save files, which
are read at server startup. These binary files also contain dynamic data that
the server has learned while operational. IPserver can be configured to
create the save files periodically, and always creates them as it shuts down.
You can also cause the files to be created immediately when using
IPmanager.
If both textual configuration files and save files exist, IPserver uses the files
with the latest time stamp. So, regardless of the method of configuration, the
server uses a combination of the manually entered configuration, and the
dynamically learned information. For example, assume you install your
IPserver, and fully configure it via the textual configuration files. The first time
you exit the server, the binary save files are created. These files will contain
both your static configuration, and the dynamic information the server learned
from the network and other Shadow IPservers. When the server is restarted,
the binary save files will have more recent time stamps than the text files, and
will be used to start the server.
If you want your IPserver to begin at an explicitly defined state each time it is
started, you can inhibit the dynamic creation of save files by including the
NOSAVE keyword in the NTS-SRVR.CFG file. You would also rename any
existing binary save files, as the server still looks for them even if the
NOSAVE keyword is present.
The textual configuration files are briefly described below. Sample
configuration screens using IPmanager can be found in 6.4.4, “Configuring
through IPmanager” on page 299.
NTS-SRVR.CFG
298
Contains basic system parameters, such as the server’s
IP address, subnet mask, and default router. These basic
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
parameters are required to enable management via the
IPmanager program or Java. This file is created by the
IPserver installation program. A sample file is shown in
Figure 216 on page 297.
NTS-SRVR.DHC
Contains the configuration for the DHCP service. This file
may not be present if you manage IPserver exclusively
with IPmanager. NTS-DHC-.SAV is the corresponding
binary save file.
NTS-SRVR.DNS
Contains the configuration for the DNS service.
NTS-DNS-.SAV is the corresponding binary save file.
NTS-SRVR.NBN
Contains the configuration for the NetBIOS Name Service.
This file can be used to pre-load NetBIOS names into the
NBNS database at startup. NTS-NBN-.SAV is the
corresponding binary save file.
NTS-SRVR.RUR
Contains optional user IDs and passwords to authorize
read-only or full access from the IPmanager or Java
configuration programs. This file must be edited manually.
All of the textual configuration files follow a similar format and use keywords
alone or in combination with parameters to define the configuration. In all,
over sixty different keywords can be used within these files. NTS provides full
product documentation in PDF format with its evaluation edition programs.
These documents describe each keyword in detail, and can also be found on
the CD-ROM that accompanies this book.
6.4.4 Configuring through IPmanager
The NTS IPmanager product runs on Windows 95/98/NT platforms, and is
installed separately. It can be used to manage multiple IPservers via an IP
connection through your network.
In the following sections, we will use IPmanager for Windows to administer
our Shadow IPserver as a DHCP, DDNS, and NBNS server.
After you install and launch IPmanager, a logon panel will be displayed,
prompting you for a user name and password. The names and passwords are
set by editing the NTS-SRVR.RUR file. You can specify read-only and full
access using multiple user IDs. The security features of IPserver are optional,
but we recommended their use, as well as an appropriate measure of
physical security around the server hardware. The Logon panel is shown in
Figure 217 on page 300.
Chapter 6. Expanding your network
299
Figure 217. [Shadow IPserver] Logon panel
Figure 218 shows our startup screen after connecting to the Shadow IPserver
by specifying its IP address. Once you connect to a server, its IP address is
saved and can be selected from the pull-down list near the top-left of the
window.
Figure 218. [Shadow IPserver] Startup screen
First, we’ll look at some global configuration by selecting
[Options—Server Configuration] from the menu bar. This displays the
Server Configuration window, as shown in Figure 219 on page 301.
300
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 219. [Shadow IPserver] Server Configuration window
The Server Configuration window is used to set global parameters for the
server. The Configuration page, shown in Figure 219 on page 301, is the first
of nine pages in the window. The Help button at the bottom of each page
Chapter 6. Expanding your network
301
brings up comprehensive contextual information. The other pages in the
Server Configuration window are described in Table 13.
Table 13. Shadow IPserver configuration pages
Page
Description
Configuration
Items specified here include the server IP address, mask and
gateway information.
Services
This page allows you to selectively enable or disable DHCP,
DNS, or NBNS services individually, without disturbing their
configuration. Logging options are also specified here.
DHCP
These pages specify global parameters for each service. If
backup or co-servers are being used, they are identified on
these pages. Parameters controlling the interaction between
this server and its peers or backups are specified here as
well.
DNS
NBNS
SNMP
Simple Network Management Protocol community names
are specified here.
Information
The build date of the IPserver software is displayed here,
along with data on capacity and resource usage.
NIC
Network adapter information such as IO address and
memory window can be seen here. Counters for packets sent
and received are also displayed.
Time
An address for a Network Time Protocol server can be
entered here. If your using backup servers or co-servers, we
recommend the use of this feature to keep all of your Shadow
IPservers’ clocks in sync.
6.4.5 DHCP configuration
IPserver lets you define multiple option sets that can contain one or many
DHCP configuration parameters. From the DHCP menu, select
[Option Sets]. A window showing the currently configured option sets will be
displayed. Figure 220 on page 303 shows the three option sets we created for
our small test networks. They are named global, short lease, and long lease.
302
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 220. [Shadow IPserver] DHCP option sets
A new option set can be added by selecting [Edit—Add] from the menu bar,
by pressing Ctrl-A, or by pressing a button on the tool bar. You can edit an
existing item by double-clicking on it, or by clicking once and selecting
[Edit—Edit] from the menu bar, by pressing Ctrl-E, or by using the Edit
button on the tool bar. The right mouse button can also be used to access
functions.There are a variety of ways to access functions in IPmanager
throughout the program.
In Figure 221 on page 304, the DHCP options contained in the global option
set are shown.
Chapter 6. Expanding your network
303
Figure 221. [Shadow IPserver] Editing DHCP Option Sets
Now, add an additional DHCP option for NetBIOS Node Type by clicking
anywhere in the options list, and then selecting the Add button. This causes
the Add/Change Options dialog to be displayed, as shown in Figure 222 on
page 305.
304
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 222. [Shadow IPserver] Add/Change Options
Select option 46 from the list of available options, and click on the radio
button for a Node Type of H Node. Pressing the Add button adds your new
option to the list. This window is also used to change the value of existing
options. Pressing the Ok button will close the window and take you back to
the prior screen.
In this scenario, you are using the global option set to contain configuration
parameters that are appropriate for a large number of clients. Also defined
are two other option sets, long lease is for our desktop clients, who rarely
move, while the short lease option set defines much shorter lease durations.
The shorter lease times are well suited to nomadic laptop users.
Chapter 6. Expanding your network
305
Figure 223. [Shadow IPserver] Long lease Option Set
Figure 223 on page 306 shows that the long lease option set specifies values
for options 51, 58 and 59. Shadow IPserver supports specification of the
three options pertaining to lease times individually. For the long lease option
set, the total lease time has been set to 360000 seconds (100 hours), while
the renewal interval is much shorter at 86,400 seconds (24 hours). The
DHCP defaults are deviated for good reasons. The liberal lease time allows
clients to run without access to a DHCP server for several days if necessary.
The relatively short renewal time (option 58) ensures that any changes made
to client configuration options will be distributed promptly.
At the bottom of Figure 223, the Option Sets list box is shown (it is empty).
IPserver allows you to refer to an option set within another option set. To do
this, click within the lower list box, then select the Add button. In
environments where the client population is very diverse you might find it
helpful to employ a greater degree of granularity in your use of option sets.
For example, you might have an option set called west coast fire wall that
contains just a single DHCP option. You can then include this option set
within other option sets instead of entering the dotted decimal address in
each option set that needs it.
306
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Duplication of options within these nested option sets is supported, however,
we recommend that you fully understand the priority algorithm IPserver
applies in such cases.
6.4.6 DHCP pools
In IPserver, DHCP pools are used to identify the IP addresses offered to
clients. They can be set up as a continuous range, a range with exclusions, or
as a simple list of available or unavailable IP addresses. Additionally, address
pools can be further defined by Client and Vendor Class options that are sent
by some clients.
Selecting [DHCP—Pools] from the menu bar will display a list of currently
defined DHCP pools, as shown in Figure 224 on page 307.
Figure 224. [Shadow IPserver] DHCP Pools
For our test networks, we are serving DHCP clients in two locations, Armonk
and Austin. Additionally, we differentiate between desktop and laptop users in
both locations. Double-clicking on the listing for armonk-laptops allows us to
view or edit the pool, as shown in Figure 225 on page 308.
Chapter 6. Expanding your network
307
Figure 225. [Shadow IPserver] Editing a DHCP Pool
At the top the screen in Figure 225 are several entry fields and pull down lists
that are used to name and control the pool. The contextual help does a good
job of describing the function of each of these items.
The lower portion of the screen in Figure 225 contains four list boxes. To add
or change the contents, click anywhere within the list box you’d like to
change, then click on the appropriate button. The function of the list boxes is
described in Table 14 on page 309.
308
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Table 14. DHCP pool configuration items
IP source ranges
Indicates the valid source subnet for these clients. This value
is calculated automatically, but can be overridden. This can
be useful in environments where the use of switching, VLANs
or ELANs results in multiple IP subnets within the same MAC
layer broadcast domain.
IP address ranges
IP addresses can be listed individually or in ranges, with any
mix of inclusion or exclusion.
Options
Individual DHCP options that are used for this pool of
addresses.
Option Sets
DHCP options that are defined using the Option Sets
functions in IPserver.
As mentioned above, we defined two DHCP pools for each of our locations
(Armonk and Austin), one pool for desktop clients and another for laptop
clients. The desktop pools use the long-lease option sets, while the laptop
pools use the short-lease option sets.
Our goal in doing this was to be able to efficiently administer our limited
number of IP addresses by assigning a much shorter lease interval to laptop
based clients. To implement this plan, we made use of a DHCP option called
User Class ID (option 77). However, we found that the Microsoft TCP/IP stack
does not support this DHCP option. To get around this problem, we installed
the IPclient software product from Network TeleSystems on each laptop. The
IPclient software enhances but does not replace the Windows TCP/IP stack.
The DHCP client in our OS/2 systems with TCP/IP Version 4.21 already
supported DHCP option 77, as did our AIX clients.
We were then able to configure all of our laptop clients to send the string
nomads as option 77 within their DHCP requests. We configured our DHCP
pools that support our laptops by placing the string nomads in the User Class
ID entry field, as seen at the top right of the screen in Figure 225 on page
308. These actions resulted in our Shadow IPserver being able to configure
our laptop clients differently than our desktop clients.
The Vendor Class ID option (66) can be implemented in a similar fashion, and
is also shown in Figure 225 on page 308.
6.4.7 DHCP configurations
Selecting [DHCP => Configurations] from the menu bar will show current
and recently configured DHCP clients, as seen in Figure 226 on page 310.
Chapter 6. Expanding your network
309
Figure 226. [Shadow IPserver] DHCP Configurations list
From this window, you can display details for a client’s configuration by double
clicking on the desired entry. The details provided for each client include the
incoming options and the list of requested options. Also listed are options the
server offered to the client, even if the client did not successfully complete the
configuration process. This type of information is helpful to have, especially in
complex installations.
Individual configurations can also be used to setup reserved addresses for a
client. This is done when you need to ensure that a client receives the same
address each time it obtains a lease.
6.4.8 Shadow IPserver DNS
In our test network, Shadow IPserver was also configured to provide static
and dynamic DNS services. Selecting [DNS => DNS Entries] from the menu
bar on IPmanager displays a dynamic list of DNS information, as seen in
Figure 227 on page 311.
310
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 227. [Shadow IPserver] DNS entries
The left portion of the window contains a tree display of Internet name
domains in the network. In this case, we have expanded the tree to view DNS
entries for “armonk.cooking.net”.
From this window you can display details, or edit individual entries. In Figure
227, note that the entries for three hosts (merlot, router, and buffet) are static,
while the other three (fillet, pizza, and wings) are dynamic. Static DNS entries
can be added by selecting [Edit—Add] from the menu bar. This brings up the
Add DNS record dialog, as shown in Figure 228 on page 312.
Chapter 6. Expanding your network
311
Figure 228. [Shadow IPserver] Add a DNS Record
In this case we are adding a type A DNS record for IP address 192.168.6.14,
a host named gumbo. Note that other DNS record types can be added as
well, as we did with our merlot host. Merlot’s TXT and HINFO records can be
seen in Figure 227 on page 311.
6.4.9 NetBIOS name services
The Shadow IPserver functions as an RFC1001/1002 NetBIOS Name Server,
as well as a NetBIOS Datagram Distributor (although some clients do not
exploit the latter function). We also used IPserver to provide Microsoft WINS
resolution to our clients. The NBNS configuration we used is shown in Figure
229 on page 313.
312
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 229. [Shadow IPserver] NBNS Configuration
The NBNS function in Shadow IPserver learns of NetBIOS hosts on the
network dynamically, as shown in Figure 230 on page 314.
Chapter 6. Expanding your network
313
Figure 230. [Shadow IPserver] NBNS database display
This display lists the NetBIOS names, their type (group or unique) and the
broadcast configuration for the host. From this window you can also add static
entries for NBNS resolution by selecting [Edit—Add] from the menu bar.
6.4.10 Backing up your IPserver configuration
Of course, once you have fully integrated IPserver into your network you will
want to protect your configuration efforts. If the server’s Save Files function is
enabled, as shown in Figure 219 on page 301, the server will create the files
each time it shuts down. So, one method of accomplishing a backup is to
simply shut down the server, and copy the binary save files to a diskette with
a DOS command:
COPY C:\SHADOW\*.SAV A:\
You can also use IPmanager to import the server’s configuration and dynamic
data as shown in Figure 231 on page 315. This will produce textual
configuration files that can be restored to the server at a later date. (Keep in
mind that *.SAV files with a newer time stamp will override older configuration
files, as described in Figure 6.4.3 on page 298.)
314
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 231. [IShadow IPserver] Import binary save files
Chapter 6. Expanding your network
315
316
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 7. Mobile users
If you're planning on supporting remote users or offices, this chapter will help
explain PPP with dial-up or ISDN connections. We'll also demonstrate how to
automatically assign printers and display welcome messages as users roam
from office to office.
7.1 PPP dial-up
PPP servers provide dial-up access to a TCP/IP network over modem, ISDN,
or similar connections for remote users working away from the office.
Sometimes these servers also link remote branch offices to the central office.
Unfortunately, most PPP servers must be managed separately, and, while
they can provide IP addresses to clients dynamically, they usually supply
those addresses out of separate, static tables maintained at each PPP server.
Consequently, they do not integrate well with DHCP, and they can impose an
extra burden on network managers.
However, IBM PPP services are included in OS/2 Warp Server for e-business
and can access a DHCP server to obtain addresses for remote users. (See
Figure 232 on page 318 and Figure 233 on page 319.) DDNS services may
also be used. By shifting the burden for address assignment back to the
DHCP server(s), the IBM PPP server can prove a lot easier to administer.
© Copyright IBM Corp. 1998, 2000
317
Figure 232. [Warp Server] PPP Server and DHCP
7.1.1 Prerequisites
To install an IBM PPP server on your network, you will need the following:
1. A system with OS/2 Warp Server for e-business and the Remote Access
Services component installed. The server should have at least one dial-up
connection available (such as a modem) for testing purposes and should
also be connected to your TCP/IP network.
2. For OS/2 Warp Server Version 4, the IBM PPP Server software can be
downloaded from the IBM Software Choice Web site at:
http://www.software.ibm.com/swchoice
You will need a Software Choice or Passport Advantage subscription,
available from your IBM software dealer, to access this software.
3. If you plan to use TCP/IP 4.1 with the PPP Server, you should also obtain
at least two fixes: APARs IC21069 and IC21116. These fixes are available
from IBM Software Support on the Internet at:
http://ps.software.ibm.com
318
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
4. A DHCP server on your network, such as OS/2 Warp Server’s TCP/IP 4.21
or 4.1 DHCP server. (In fact, the PPP Server can reside on the same
system as the DHCP server.)
5. A DDNS server if you wish to take advantage of dynamic name services.
(Optional.)
6. A remote client with a PPP dialer for testing purposes.
Figure 233. [Warp Server] PPP Server and DHCP: Example
The IBM PPP Server (formerly called LAN Distance) works with many, but not
all, network adapters supported by OS/2 Warp Server. Choose the network
adapter carefully for your PPP server. A list of supported adapters can be
found on the Internet at the OS/2 Device Driver Pak On-Line at:
http://service.software.ibm.com/os2ddpak/index.htm
With the correct communications adapter, the IBM PPP Server can support
up to 128 dial-up ports per server using products such as NETAnywhere. For
more information on NETAnywhere, please visit:
http://vrcomm.com/html/vcpos2.html
Chapter 7. Mobile users
319
7.1.2 Step-by-step procedure
Here are the major steps required to set up the IBM PPP Server with DHCP:
1. Install the IBM PPP server software. Simply follow the normal procedure
described in the documentation.
Figure 234. [Warp Server] TCP/IP added to PPP Logical Adapter
2. Verify that the TCP/IP protocol is bound to both adapters. Using MPTS,
make sure that the TCP/IP protocol is listed under both network adapters.
0 - IBM TCP/IP should be listed under the regular LAN adapter and 1 IBM TCP/IP should be listed under the LAN Distance Logical Adapter. See
Figure 234.
3. Configure IP addresses for both adapters. Note that these addresses need
not be on the same subnet. Although DHCP can be used to assign these
addresses, you may wish to stick with static addresses until your other
testing has been completed.
4. Turn on IP forwarding. TCP/IP traffic must be allowed to flow from the LAN
Distance Logical Adapter to your regular LAN adapter and onto the
network. Open the TCP/IP Configuration notebook and check the IP
320
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Forwarding checkbox. See Figure 235.
Figure 235. [Warp Server] IP Forwarding enabled
5. Edit WCLLOCAL.INI to turn on DHCP. In the \WAL directory, a key
configuration file for the PPP Server, WCLLOCAL.INI, can be found. Using
any text editor, such as the OS/2 System Editor, find the [PPP] section and
change the OBTAINIPADDR line so that it reads:
OBTAINIPADDR=DHCP
(You may also wish to activate DDNS.)
6. Define an address pool for PPP clients on the DHCP server. The subnet
containing the dial-up PPP clients should have enough free addresses
available.
Chapter 7. Mobile users
321
7. Define a PPP user. Start the IBM Remote Access program. Using this
graphical program, define at least one PPP user for testing.
8. Add the PPP Server to the Startup folder. Drag a shadow of the IBM
Remote Access program icon to the Startup folder to make sure that the
PPP Server always starts up every time the server reboots.
9. Test the dial-up connection. Your dial-up PPP client should be able to
connect to the server, get an address assignment and communicate with
other systems using the user ID and password you defined earlier.
External modems on both ends can prove helpful since they provide status
indicators which can be used to monitor the progress of connections.
The online PPP Server reference guide includes more information on
troubleshooting your dial-up connections. In particular, you should pay close
attention to the routing tables and subnets you have implemented on your
network, since most common errors occur when dial-up clients do not have
proper routing to other systems on the network.
7.2 Roaming users
While we've seen how people can work at home through a dial-up PPP
connection and get access to the office TCP/IP network, it isn't the only place
where dynamic network access may be needed. Many companies have
several different locations and workers who travel may need to connect from a
remote office. It's important that they feel at home, getting not only IP address
information but also access to local network printers and any other resources
particular to that specific office. That way, travelers can work as quickly and
easily as possible.
Sales people, service personnel, executives, and many other professionals
must work in several different locations from time to time, and many of them
rely heavily on their notebook computers. Hotelling is also becoming
commonplace, with telephone systems and computing technologies that
allow employees to sit at any desk in the office. A dynamic IP network can
help respond to these changing business needs. This section focuses on
extending DHCP to help assign local printers, display a local message of the
day, and otherwise provide simplified access to the network wherever a
roaming user may be.
Additionally, you may require lease differentiation between laptop and
desktop computers, wishing to give laptop computers shorter lease times
than desktop computers. See 6.4.6, “DHCP pools” on page 307 for a
description of how to accomplish this with NTS Shadow IP Server.
322
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
For this section, we tested OS/2 Warp 4 on an IBM ThinkPad 760ED
notebook computer with the IBM PCMCIA Token-Ring Adapter. However, you
may adapt these scenarios to your own particular software, equipment, and
network. In fact, since almost every company has roaming users of one sort
or another, we believe most readers ought to study this section carefully.
7.2.1 Prerequisites
To provide services for roaming users, you will first need access to the
following resources:
1. A DHCP server (preferably with DDNS), such as OS/2 Warp Server. We
recommend you set a relatively short lease time (such as three minutes;
three minutes is the shortest lease time allowed by the DHCP server) for
testing purposes.
2. A sample client system, with basic DHCP capabilities (and, if necessary,
file/print services with TCPBEUI) working correctly.
3. Access to at least one other location, whether it's across the hall, another
floor, or in a different building. This location may be on a different subnet.
4. A Web server, such as Lotus Domino, for message of the day service.
5. A network printer, on a OS/2 Warp Server or NT domain, or as a TCP/IP
(LPD) printer.
7.2.2 Automatic connection to file/print domain
If you primarily use traditional file and print services, reserving TCP/IP for
Web browsers and other applications, then you can use extended DHCP
options to pass information on the local file/print domain to a roaming client.
7.2.2.1 Assumptions
For now, we assume:
• A roaming user visits at least two different offices, floors, or other separate
file/print domains with a notebook computer running OS/2 Warp 4.
• The roaming user has an account on each domain (even a GUEST
account) with default printer assignments.
• The domain controller and the client can use TCPBEUI to communicate.
• For simplicity, the domain controller is doubling as the print server, and the
client does not need to use TCPBEUI to communicate with any other
systems. In other words, only one TCPBEUI broadcast address is needed
by the client.
• A NetBIOS name server, such as Shadow IPserver, is not available.
Chapter 7. Mobile users
323
Both the domain name and the IP address of the domain controller may vary
from location to location.
7.2.2.2 DHCP server setup
DHCP provides some standard options (44 to 47) for passing
NetBIOS-related information to clients. However, none of these options seem
appropriate given our assumptions. For example, there’s no explicit option to
provide the domain name to our client. While it’s possible to use a published
DHCP option for something other than its specified use, we certainly don’t
recommend the practice.
Instead, there are at least two ways to extend DHCP to provide this custom
information to each client. One way is to use Option 43 (Vendor Specific
Information). Perhaps the best way, though, is to use any of the options
already set aside for custom use, options 128 through 254.
Although not universally accepted, many platforms have settled on specific
purposes for some of the DHCP options in this range. OS/2 Warp, for
example, may use options 200 to 208 to receive various extra TCP/IP
parameters, such as the default SOCKS server address (Option 205).
Consequently, these should be avoided as well when choosing your own
custom options.
324
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 236. [Warp Server] Adding class definition to DHCP server
We need to add two extra custom options to the DHCP server. The following
two options can be defined:
Option 150
Option 151
Domain Name
TCPBEUI Broadcast Address
Chapter 7. Mobile users
325
Figure 237. [Warp Server] Adding option 150 to class IBMWARP_V3.1
For our first test office, the domain name is ARMONK, and the TCPBEUI
broadcast address is merlot.armonk.cooking.net. Since these are not
reserved DHCP options, they shouldn’t bother any other DHCP clients; so,
you can safely add them to any subnet you wish. Try adding them to your
DHCP server for a specific subnet, client classification, or other portion of
your network where they would apply.
In Figure 236 on page 325, we define a new class, IBMWARP_V3.1, for the
entire subnet, armonk.cooking.net. The class name, IBMWARP_V3.1, is
found in the DHCPCD.CFG file on an OS/2 Warp client, and it’s normally the
default value for both options 60 and 77. (You can edit DHCPCD.CFG on the
client if you wish to change this value.) This class name is communicated to
the DHCP server at the time the client requests its address assignments. In
this example, listing options 150 and 151 under a specific class name
prevents other clients, even within the same subnet, from receiving these
extra option values.
Figure 237 and Figure 238 on page 327 show how one of these extra options,
the LAN Domain Name, is defined and its value, ARMONK, set.
326
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 238. [Warp Server] Setting ARMONK as DHCP Option 150 Value
Finally, in Figure 239 on page 328, all configuration is complete. With our new
class highlighted, IBMWARP_V3.1, the options that will be received by
members of this class are listed with their value on the right. In particular,
option 150 will provide the string ARMONK, and option 151 will provide the
fully qualified domain name for the domain controller, the string
merlot.armonk.cooking.net. In fact, merlot.armonk.cooking.net is registered
with a dynamic DNS server, and its numeric IP address assignment could
change; so, it’s particularly important that a name server be provided to the
client as one of the options to help properly resolve this TCP/IP address.
Although options 44, 45, and 46 are defined for this entire subnet, they will
not be used by this client. We assume that a NetBIOS name server is not
available.
Chapter 7. Mobile users
327
Figure 239. [Warp Server] Option values for class IBMWARP_V.3.1
7.2.2.3 Client setup
Our client (OS/2 Warp 4 in this example) has been previously set up to get its
IP address assignment using DHCP and to use TCPBEUI as the protocol for
its file/print client software. Now it’s up to OS/2 Warp 4 to do something with
these two new DHCP values.
In OS/2 Warp 4, the file DHCPCD.CFG controls the DHCP client
configuration. This file is located in the directory pointed to by the ETC
environment variable, as set in the CONFIG.SYS file. Normally, but not
always, this directory is X:\MPTN\ETC (where X: is the OS/2 Warp boot
drive). You can edit this file with any text editor, such as the OS/2 Warp
System Editor.
Part of the DHCPCD.CFG file describes the DHCP options the client should
handle (and how they should be handled). That section appears below:
#option
#option
#option
#option
#option
[...]
328
9
71
200
201
202
exec
exec
exec
exec
exec
"dhcpibm.cmd
"dhcpibm.cmd
"dhcpibm.cmd
"dhcpibm.cmd
"dhcpibm.cmd
9 %s"
71 %s"
200 %s"
201 %s"
202 %s"
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
#
#
#
#
#
LPR Server
Default NewsReader/2
Default LPR Printer
Gopher Server
Default WWW Home Page
The # symbol indicates that the option is commented out and deactivated. We
need to add options 150 and 151 and process them appropriately. So, in this
section, add the following lines (without the # symbol in front):
option 150 exec "myoffice.cmd 150 %s"
option 151 exec "myoffice.cmd 151 %s"
# Domain Name
# TCPBEUI Broadcast Address
By adding these lines to the DHCPCD.CFG file, we’re instructing OS/2 Warp
4 to obtain these extra option values (150 and 151) from the server and to
pass them to a REXX script called MYOFFICE.CMD, which you can create.
MYOFFICE.CMD will be run twice, once for each option. Unless the directory
is specified (for example, C:\BATCH\MYOFFICE.CMD), the PATH (as set in
CONFIG.SYS) must contain the directory where MYOFFICE.CMD is located.
DHCPIBM.CMD is provided as a sample REXX script to show how DHCP
options can be processed. You can find DHCPIBM.CMD in the \MPTN\BIN
directory.
Here’s how MYOFFICE.CMD might look, using DHCPIBM.CMD as a basis:
/* MYOFFICE.CMD
* by Timothy Sipples
* MYOFFICE <option tag> <option data>
* Return values: 0=successful, 1=input error, -1=system error
*/
/* add RexxUtil functions */
call RxFuncAdd 'SysFileDelete', 'RexxUtil', 'SysFileDelete'
return_code = 0
bcst_file = "C:\IBMCOM\RFCBCST.LST"
/* Drive C; change if needed */
/* Get option number and data from the command line */
parse arg option_tag option_data
if ((option_tag = '') | (option_data = '')) then do
return_code = 1
/* wrong number of arguments */
exit return_code
end
/* Strip leading and trailing spaces */
option_tag = strip(option_tag)
option_data = strip(option_data)
Figure 240. [OS/2 Warp] MYOFFICE.CMD (Part 1 of 3)
Chapter 7. Mobile users
329
/* Instantiate the option */
select
when (option_tag = 150) then do
/* Domain Name */
data = hex2ascii_string(option_data)
if (data \= '') then return_code = change_domain(data)
else return_code = 1
/* improper Domain Name */
end
when (option_tag = 151) then do
/* TCPBEUI B'cast Address */
data = hex2ascii_string(option_data)
if (data \= '') then do
dummy=SysFileDelete(bcst_file)
return_code = lineout(bcst_file,data)
dummy=stream(bcst_file,"C","CLOSE")
"@rfcaddr"
end
else return_code = 1
/* improper B'cast Address */
end
otherwise
return_code = 1
end
exit return_code
/* hex2ascii_string (option_data)
* Take a hex string in the form "nn nn nn ...", verify it's a valid
* hex string, and convert it to an ASCII character string.
* Returns the string if the data is valid, a null string if not.
*/
hex2ascii_string: procedure
parse arg 'hex' '"' data '"'
if (\ DATATYPE(data, X)) then return ""
data = strip(data, 'B', '09'x)
return X2C(data)
/* change_domain (domain_name)
* Take the domain_name and record it in IBMLAN.INI
*/
change_domain: procedure
parse arg domain_name
lanini_file = "C:\IBMLAN\IBMLAN.INI"
Figure 241. [OS/2 Warp] MYOFFICE.CMD (Part 2 of 3)
330
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/* Change if needed */
temp_file = "C:\IBMLAN\IBMLAN.TMP"
/* Change if needed */
dummy = SysFileDelete(temp_file)
dummy = stream(lanini_file,"C","OPEN READ")
dummy = stream(temp_file,"C","OPEN WRITE")
do while lines(lanini_file)
temp_line = linein(lanini_file)
if left(translate(strip(temp_line)),6) = "DOMAIN" then
temp_line = " Domain = "||domain_name
dummy = lineout(temp_file,temp_line)
end
dummy=stream(lanini_file,"C","CLOSE")
dummy=stream(temp_file,"C","CLOSE")
dummy=SysFileDelete(lanini_file)
"@copy "temp_file" "lanini_file
return 0
Figure 242. [OS/2 Warp] MYOFFICE.CMD (Part 3 of 3)
When DHCP option 150 is received by MYOFFICE.CMD, the domain name is
extracted and recorded in the IBMLAN.INI file. When DHCP option 151 is
received, MYOFFICE.CMD extracts the broadcast address and creates a new
RFCBCST.LST file containing that address. Then RFCADDR runs and
rereads the broadcast and names files (RFCBCST.LST and
RFCNAMES.LST) for TCPBEUI.
When the desktop loads, STARTUP.CMD works in partnership with
MYOFFICE.CMD. Figure 243 on page 332 shows how STARTUP.CMD may
look.
Chapter 7. Mobile users
331
/* STARTUP.CMD
* by Timothy F. Sipples
* Waits for MYOFFICE.CMD to rewrite IBMLAN.INI then proceeds
* to start requester and prompt for logon.
*/
lanini_file = "C:\IBMLAN\IBMLAN.INI"
/* Change if needed */
call RxFuncAdd "SysFileTree","RexxUtil","SysFileTree"
call RxFuncAdd "SysSleep","RexxUtil","SysSleep"
"@cls"
say "Waiting to receive your local domain name."
say "One moment, please..."
do forever
dummy = SysFileTree(lanini_file,"info","T")
curr_date = date("O")
curr_time = time("M")
parse value info.1 with f_yy "/" f_mo "/" f_dd "/" f_hh "/" f_mm " " .
file_date = f_yy||"/"||f_mo||"/"||f_dd
file_time = f_hh*60 + f_mm
if curr_date = file_date then
if (curr_time = file_time) | (curr_time = file_time+1) then leave
if (curr_time = 0) & (file_time = 60*23+59) then do
/* Insert midnight rollover handling here. */
end
call SysSleep 10
end
"@START /N /PM LOGON /V:DOMAIN"
"EXIT"
Figure 243. [OS/2 Warp] STARTUP.CMD
STARTUP.CMD waits for MYOFFICE.CMD to rewrite IBMLAN.INI with the
correct default domain name. Then STARTUP.CMD starts LOGON, which first
starts LAN Requester. The user can then log onto the local domain to obtain
default printer assignments and other LAN resources.
332
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
STARTUP.CMD knows that MYOFFICE.CMD has finished with IBMLAN.INI by
comparing the file date and time with the current date and time. If these two
times differ by no more than one minute, MYOFFICE.CMD has likely done its
work, and the LAN Requester starts. If not, STARTUP.CMD keeps checking
every 10 seconds. Since MYOFFICE.CMD could take up to a minute or more
to get the extra DHCP options, STARTUP.CMD warns the user with a
message, One moment, please....
7.2.2.4 Technical caveats
You should view MYOFFICE.CMD and STARTUP.CMD as examples to
illustrate the capabilities of DHCP in helping to support mobile users. Many
potential problems exist in both batch files. We would like to draw your
attention to a few of these issues:
• MYOFFICE.CMD runs with every lease renewal, once for each of the two
new DHCP options. If the lease period is short, then a lot of background
processing can take place, impacting performance. Also, tasks that require
user input (such as LOGON) should not be started from MYOFFICE.CMD
or any equivalent. If the lease time is two minutes, for example, then the
user would be prompted every two minutes to log onto the network. Use
the DHCP option handler to record option values in INI files, configuration
files, and other locations. Use other batch files (such as STARTUP.CMD or
a program started from the Startup folder) to act on these option values.
• We believe that it’s appropriate to delete RFCBCST.LST and rewrite the
file with each renewal. A more sophisticated version of MYOFFICE.CMD
should perhaps modify entries in RFCBCST.LST and RFCNAMES.LST
more carefully, allowing the client to communicate with other systems
using TCPBEUI. Or option 151 could be extended to provide a list of many
names and/or addresses, a poor man’s NetBIOS name server.
• Comparing the current date and time with the file date and time in
STARTUP.CMD is not the most elegant method of communication between
MYOFFICE.CMD and STARTUP.CMD. You should consider using REXX
queues, for example, if you need to have the DHCP option handler alert
another one of your batch files that an event (receipt of the DHCP option)
has occurred. Remember, however, that MYOFFICE.CMD runs for each
option and for each lease renewal, so you should be careful not to keep
adding to a REXX queue indefinitely to avoid running out of memory. The
best approach might be to use a separate queue for each notification and
to clean out each queue before use. (See 7.2.3, “Netscape Message of the
Day service” on page 334, for an example.)
• Error handling could be substantially improved in these batch files, to
catch disk full errors, for example.
Chapter 7. Mobile users
333
• We used the fully qualified domain name merlot.armonk.cooking.net in
option 151. We discovered that enabling the OS/2 Warp SOCKS capability
in the TCP/IP stack (SOCKS_FLAG on in the \MPTN\ETC\SOCKS.ENV
file) disrupted name resolution with the fully qualified name if the SOCKS
server was unreachable. You may wish to use the hostname only (merlot
in this example) if you are also using DHCP to pass the standard domain
name option and if the hostname alone can be resolved.
• In these batch files there’s no guarantee that option 151 handling will be
complete before the user attempts a logon. STARTUP.CMD could be
enhanced to also check that RFCBCST.LST has been updated and
RFCADDR has been run before starting LOGON. Again, REXX queues
may help.
• Be aware that using the graphical setup programs in OS/2 Warp (such as
the TCP/IP Configuration Notebook or the Multiprotocol Transport
Services Setup) can rewrite IBMLAN.INI, DHCPCD.CFG, and other critical
files. Make sure that any changes you make to these files in order to
support enhanced DHCP option handling are not reversed or modified
when you take advantage of the graphical setup programs.
• We do not necessarily recommend placing all the REXX programming
code for enhanced DHCP option handling on every client. Instead, you
may wish to pass a single DHCP option to a client containing the latest
version of the REXX script(s) to be run. This bootstrap method can help
make managing clients easier.
• Handling TCP/IP printers is also quite easy, and some example REXX
scripting appears in DHCPIBM.CMD for the relevant DHCP options (9 and
200).
7.2.3 Netscape Message of the Day service
Based on the file/print example, it should also be possible to provide a
message of the day service to roaming clients and improve upon the methods
used previously. The DHCP server will pass a Web address (URL) to the
client as one of the options, and the client will start Netscape at bootup to
display this message.
7.2.3.1 DHCP server setup
Appendix D, “DHCP options (RFC 2132)” on page 457, lists the DHCP
options and their purposes. While not officially part of the standard, DHCP
option 114 is listed as "Universal Resource Locator (URL)," and it seems to
be the most appropriate to pass a Web address to a client. Or, if you prefer,
you may use one of the DHCP options strictly reserved for your own use,
options 128 through 254.
334
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
We added Option 114 with a value of:
http://merlot.armonk.cooking.net/siteinfo.html
to the same class and subnet as shown in 7.2.2.2, “DHCP server setup” on
page 324.
7.2.3.2 Web server setup
We opted to combine our DHCP server with our Lotus Domino Web server on
the same system, merlot.armonk.cooking.net. However, you may wish to have
one central Web server with all the message of the day Web pages available.
For example, you might arrange your Web server to provide access to the
following URLs:
http://www.siteinfo.megahuge.com/chicago
http://www.siteinfo.megahuge.com/newyork
http://www.siteinfo.megahuge.com/london
http://www.siteinfo.megahuge.com/peoria
depending on the location (class, subnet, and so forth) of the client.
Assuming there’s enough capacity, managing one central Web server can be
easier.
However, in this simple example, our DHCP server doubles as our Web
server. We placed a test Web page (siteinfo.html) in the root directory of this
Web server and verified that the Web page could be loaded across the
TCP/IP network using a standard Web browser.
7.2.3.3 Client setup
Again, in this example, our OS/2 Warp 4 client correctly obtains its DHCP
address assignment information from merlot.armonk.cooking.net. Also,
Netscape for OS/2 Warp can properly load siteinfo.html from merlot. Now it’s
up to the client to automatically display that Web page at startup, immediately
after receiving the DHCP option.
For this example, you should edit DHCPCD.CFG and add the following line:
option 114 exec "$mywebmsg.cmd %s"
# Message of the day (URL)
MYWEBMSG.CMD is a REXX script located in some directory in the PATH.
Unlike MYOFFICE.CMD in the previous example, which handled two DHCP
options, MYWEBMSG.CMD just handles the contents of the single DHCP
option (114) passed to the client.
The $ means to wait 30 seconds before starting the REXX script to allow
enough time for REXXINIT.DLL to start. REXXINIT.DLL allows REXX scripts
Chapter 7. Mobile users
335
to properly function. Alternatively, a & symbol can be used to specify a
program that requires OS/2 Warp’s Presentation Manager. Without the $ or &
symbols, OS/2 Warp’s DHCP client software will attempt to run the program
immediately upon receiving the option at bootup, possibly before the desktop
is ready.
We made some significant improvements to the DHCP option handler,
MYWEBMSG.CMD, in order to demonstrate a better way to handle delivery of
DHCP options dynamically. MYWEBMSG.CMD obtains the Web address and
places it on a REXX queue called MYWEBMSG instead of recording it in a
file.
Figure 244 on page 337 displays MYWEBMSG.CMD.
336
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/* MYWEBMSG.CMD
* by Timothy Sipples
* MYWEBMSG <option data>
* Return values: 0=successful, 1=input error, -1=system error
* Designed to handle DHCP option 114 and to place the
* contents of the option (web address) into a REXX queue.
return_code = 0
q_name = "MYWEBMSG"
*/
/* Get option data from the command line */
parse arg option_data
if option_data = '' then do
return_code = 1
/* No data returned */
exit return_code
end
/* Strip leading and trailing spaces */
option_data=strip(option_data)
data=hex2ascii_string(option_data)
if data = '' then do
return_code = 1
/* String is blank */
exit return_code
end
/* Put data on REXX queue */
call RXQUEUE "Delete",q_name
call RXQUEUE "Create",q_name
call RXQUEUE "Set",q_name
push data
exit return_code
/* hex2ascii_string (option_data)
* Take a hex string in the form "nn nn nn ...", verify it's a valid
* hex string, and convert it to an ASCII character string.
* Returns the string if the data is valid, a null string if not.
hex2ascii_string: procedure
parse arg 'hex' '"' data '"'
if (\ DATATYPE(data, X)) then return ""
data = strip(data,'B','09'x)
return X2C(data)
Figure 244. [OS/2 Warp] MYWEBMSG.CMD
Chapter 7. Mobile users
337
Note that MYWEBMSG.CMD deletes the REXX queue each time it runs, with
each lease renewal.
Figure 245. [OS/2 Warp] Program object in startup folder (Page 1)
Figure 246. [OS/2 Warp] Program object in startup folder (Page 2)
338
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The companion REXX script, SEEMYMSG.CMD, starts from a program object
in the Startup folder. (See Figure 245 and Figure 246 on page 338.)
SEEMYMSG.CMD waits indefinitely until the Web address appears in the
REXX queue. When it appears, SEEMYMSG.CMD starts Netscape in kiosk
mode ( -k option). In our tests, the message of the day appeared
approximately 30 seconds after bootup.
Figure 247 displays SEEMYMSG.CMD.
/* SEEMYMSG.CMD
* by Timothy Sipples
* Waits for web address to be placed on queue MYWEBMSG,
* reads it, then starts Netscape with the web address.
* Designed to launch from Startup folder.
*/
q_name = "MYWEBMSG"
address CMD
/* add RexxUtil functions */
call RxFuncAdd 'SysSleep','RexxUtil','SysSleep'
do forever
call RXQUEUE "Set",q_name
if QUEUED() > 0 then do
parse pull data
"@START /N /PM NETSCAPE -3 -k "data
exit
end
call SysSleep 5
end
Figure 247. [OS/2 Warp] SEEMYMSG.CMD
We prefer the REXX queue method of communication between the DHCP
option handler and other REXX scripts. However, both methods can be
combined when it’s necessary to update a configuration file on the client.
Of course, our simple message of the day (Figure 248 on page 340) can be
extended to include links to important company Web applications and forms,
information on support staff at a particular site, the current stock price, a map
of the building, or any other pertinent information that can help a traveler feel
at home in a new location. Kiosk mode prevents access to menu and URL
controls so that your message (and links) are the only ones available.
Chapter 7. Mobile users
339
Figure 248. [OS/2 Warp] Netscape in Kiosk mode
7.2.3.4 Technical caveats
We recognize that even MYWEBMSG.CMD and SEEMYMSG.CMD are not
perfect. Here are some technical issues that you may wish to explore further:
• SEEMYMSG.CMD runs forever (do..forever loop). You may wish to modify
this REXX script so that it only tries to obtain the Web address for a
particular length of time, perhaps displaying an error message after the
time has elapsed.
• If Netscape has been configured with a proxy server of some kind, to
access Web pages outside a firewall, then that server should be available
from every site. Proxy error messages will prevent Netscape from loading
and displaying the message of the day. If necessary you can use the -i
command line option to force Netscape to use an alternative
NETSCAPE.INI file, one which has proxy capability turned off. That INI file
may come from a file server. For example:
NETSCAPE -iP:\COMMON\COMMON.INI -3 -k http://www.megahuge.com/rome
See the Netscape README file and 7.2.4, “Custom Netscape INI files” on
page 341 for details.
Of course, you can also use Netscape Navigator as a gateway to other
applications by specifying custom Helper properties in your COMMON.INI
file, so that Java and EXE (binary) applications start automatically or start
with the selection of a menu option. In other words, you can create a
complete desktop environment and use Netscape to deliver it on demand.
• As before, the SOCKS server must be reachable from any site
(SOCKS.ENV file); otherwise OS/2 Warp 4 will have trouble resolving fully
qualified names.
• Although MYWEBMSG.CMD runs with each lease renewal, there should
be little or no performance impact since it’s not necessary to rewrite a file.
340
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Also, SEEMYMSG.CMD exits as soon as any Web address appears in the
REXX queue.
• Both REXX scripts should include more error checking, particularly in the
handling of the REXX queue.
7.2.4 Custom Netscape INI files
While it’s certainly possible to pass options to Netscape using the command
line, it may also be useful to directly access the NETSCAPE.INI file in order to
modify it. Netscape’s INI file contains many more options that control the
browser’s behavior, and it’s possible to influence that behavior based on
option values delivered by a DHCP server.
For example, roaming users may need to use different proxy or SOCKS
firewalls depending on which office (or subnet) they happen to be at. Or, your
network may be subject to change from time-to-time, with addresses for
firewalls changing periodically as your needs change. Changing Netscape
settings (particularly complex firewall options) can be tedious for every user;
so it makes sense to make these changes automatically and dynamically from
a central location.
In this example, we illustrate how a DHCP server can provide the address of
a SOCKS server to a client. A DHCP option handler, NSSOCKS.CMD,
receives this address and records it in the client’s NETSCAPE.INI file. Similar
option handlers can be used to modify proxy firewall addresses and other
Netscape settings.
7.2.4.1 DHCP server setup
By convention, IBM reserves DHCP Option 205 to provide the address of a
SOCKS server to a client. (See Appendix D.4, “Unofficial DHCP options” on
page 481.) Therefore, at your DHCP server. you should add option 205 and
its value to each subnet, class, or other part of the network where you wish
this address to take effect.
For instance, in your network you may have three firewalls:
socks1.cooking.net
socks2.cooking.net
socks3.cooking.net
and each firewall could serve a specific subnet (or group of subnets).
Therefore, users on a particular subnet should get the address of the
matching SOCKS server. In addition, if one firewall fails, users on the affected
subnet could be reassigned to one of the other firewalls by making a simple
administrative change at the DHCP server.
Chapter 7. Mobile users
341
Refer to 7.2.2.2, “DHCP server setup” on page 324, for information on how to
add a custom DHCP option value to the server.
7.2.4.2 Client setup
For this example, we again have a fully functioning OS/2 Warp 4 client with
Netscape Navigator 2.02 for OS/2 Warp. The client can properly receive
regular DHCP address assignment information and can connect to the
network.
By default, Netscape Navigator stores its settings in the file NETSCAPE.INI,
normally located in the same directory containing the other Netscape
program files, such as NETSCAPE.EXE. NETSCAPE.INI may be hidden (or
may have the system attribute set). You can remove these attributes by using
the command:
ATTRIB -R -H -S C:\NETSCAPE\NETSCAPE.INI
assuming that you have installed Netscape in the C:\NETSCAPE directory.
You may wish to change these file attributes for convenience before working
with NETSCAPE.INI. You may also wish to make a backup copy of
NETSCAPE.INI before attempting to make changes to it.
In 7.2.3.3, “Client setup” on page 335, we showed how to install a custom
DHCP option handler. For this example, we again installed an option handler
on the client, this time NSSOCKS.CMD for option 205. Add the following line
to DHCPCD.CFG:
option 205 exec "$nssocks.cmd %s"
#SOCKS firewall server address
and place the following REXX script, NSSOCKS.CMD, somewhere in the
client’s PATH (see Figure 249 and Figure 250 on page 344):
342
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/*
*
*
*
*
NSSOCKS.CMD
by Timothy Sipples
Example DHCP option handler which changes SOCKS server setting
for NETSCAPE.INI file to value specified as an argument (option 205).
Return values: 0=successful, 1=error
*/
return_code = 0
address CMD
call RxFuncAdd "SysLoadFuncs","RexxUtil","SysLoadFuncs"
call SysLoadFuncs
/* Find full path to NETSCAPE.INI */
IniFile = SysSearchPath("PATH","NETSCAPE.INI")
if IniFile = '' then do
return_code = 1
exit return_code
end
/* Get option data from the command line */
parse arg option_data
if option_data = '' then do
return_code = 1
exit return_code
end
/* Strip leading and trailing spaces */
option_data = strip(option_data)
data = hex2ascii_string(option_data)
if data = '' then do
return_code = 1
exit return_code
end
/* Modify NETSCAPE.INI with SOCKS server */
result = SysIni(IniFile,"Services","SOCKS_Server",data)
if result \= '' then do
return_code = 1
exit return_code
end
Figure 249. [OS/2 Warp] NSSOCKS.CMD (Part 1 of 2)
Chapter 7. Mobile users
343
result = SysIni(IniFile,"ProxyInformation","ProxyPref","ManualProxy")
if result \= '' then do
return_code = 1
exit return_code
end
/* Set SOCKS server port to 1080 (assumed).
* Note that 1080 (decimal) is 0438 (hex), which must be in reverse
* byte order and padded with zeroes (8 digits) for the SysIni
* function, as shown here. */
result = SysIni(IniFile,"Services","SOCKS_ServerPort",x2c('38040000'))
if result \= '' then do
return_code = 1
exit return_code
end
exit return_code
/* hex2ascii_string (option_data)
* Take a hex string in the form "nn nn nn ...", verify it's a valid
* hex string, and convert it to an ASCII character string.
* Returns the string if the data is valid, a null string if not.
*/
hex2ascii_string: procedure
parse arg 'hex' '"' data '"'
if (\ DATATYPE(data, X)) then return ''
data = strip(data,'B','09'x)
return X2C(data)
Figure 250. [OS/2 Warp] NSSOCKS.CMD (Part 2 of 2)
NSSOCKS.CMD first searches the PATH to locate NETSCAPE.INI using the
SysSearchPath function. The, the DHCP option value gets converted from its
hexadecimal representation to a text string (such as socks2.cooking.net).
Finally, NSSOCKS.CMD makes three changes to NETSCAPE.INI using the
SysIni function.
7.2.4.3 Other Netscape INI Settings
You can explore NETSCAPE.INI (or any other standard binary INI file used to
store application settings) by running REGEDIT2. REGEDIT2 ships with
recent IBM FixPaks for OS/2 Warp and OS/2 Warp Server.
344
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Note that the format of NETSCAPE.INI could change; so you should verify
that you are changing the correct settings if you upgrade your version of
Netscape Navigator.
Here’s a partial list of the Netscape settings you may wish to change using
DHCP option handlers. To change an INI file setting with REXX, use the
SysIni function:
result = SysIni(IniFile,Application,Key,Value)
NETSCAPE.INI Settings (partial list)
Application: ProxyInformation
Key: ProxyPref
Value:
Key: AutoProxyUrl
Value:
Key: FTP_Proxy
Value:
Key: FTP_ProxyPort
Value:
Key: Gopher_Proxy
Value:
Key: Gopher_ProxyPort Value:
Key: HTTPS_Proxy
Value:
Key: HTTPS_ProxyPort
Value:
Key: Http_Proxy
Value:
Key: Http_ProxyPort
Value:
Key: No_Proxy
Value:
Key: Wais_Proxy
Value:
Key: Wais_ProxyPort
Value:
Application: Services
Key: SOCKS_Server
Key: SOCKS_ServerPort
NoProxy|ManualProxy|AutoProxy
(URL string)
(URL string)
(8X)
(URL string)
(8X)
(URL string)
(8X)
(URL string)
(8X)
(string)
(URL string)
(8X)
Value: (URL string)
Value: (8X)
8X = eight digit hexadecimal, reverse byte order
7.2.4.4 Technical caveats
NSSOCKS.CMD can be enhanced or extended in various ways:
• If you start Netscape Navigator before NSSOCKS.CMD completes its
changes to NETSCAPE.INI, then you could pick up old settings. Try
adding a REXX queue signalling method to your own DHCP option handler
like the one shown in 7.2.3.3, “Client setup” on page 335. Also, once
running, Netscape Navigator may not pick up subsequent changes to
NETSCAPE.INI, which may occur with each lease renewal.
• This approach can be used with WorkSpace On-Demand, assuming your
DHCP option handler has write permission to the appropriate
NETSCAPE.INI file(s). Although NETSCAPE.INI files are stored on the file
server, and although they can then be centrally managed, you may still
prefer to manage certain Netscape settings through your DHCP server
Chapter 7. Mobile users
345
instead. (In fact, the DHCP option handler could run on the file server, in
the background, updating every user’s NETSCAPE.INI file as needed.)
7.2.5 Summary
By combining custom DHCP options, and providing sophisticated client
handling of those options, you can help provide a friendly desktop for your
roaming users no matter where they happen to be. Consider using DHCP not
just to provide a basic network connection, but also to enrich the desktop
experience and to allow people to get to work more quickly and easily.
Although we’ve shown one specific type of client working with these custom
DHCP options, you may be able to adapt these techniques for other clients,
assuming they support custom DHCP options. Having a scripting language
available, such as REXX or BASIC, can help make the task that much easier.
346
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 8. Security of DHCP and Dynamic DNS
Security is often an afterthought in building a network, but you should be very
concerned whether your private (or even public) TCP/IP network can prevent
unwanted access. In this chapter, we show how to refuse connections to the
network so that no one with a notebook computer can walk in and grab your
information (unless you want them to). One basic security issue is to make
sure that you have the latest upgrades to your systems that are depended
upon to provide network connectivity. All systems should be at the latest
upgrades, especially those that connect to the network. Manufacturers
generally respond fairly quickly when hackers discover another hole in their
software.
A good mailing list to be on for security issues is the Computer Emergency
Response Team (CERT) list. You are notified of security holes as they are
published. See http://www.cert.org to get on the mailing list, or
http://www.cert.org/advisories for the latest information.
8.1 Security trade-off
For users who want easy access and less procedure to hook up their
workstations, DHCP is very convenient for both users and administrators.
TCP/IP administrators will be relieved from the tedious work in assigning the
IP address to a new request.
For a static IP address assignment, the administrator normally assigns the
hostname to the address. It will be treated like a memo. The hostname would
be john-doe, for example. The assigned hostname is a kind of link to another
document, which will have the actual name of the owner and phone number,
and so on. However, the whole static TCP/IP itself has no security to prevent
an unauthorized access to the IP network. By checking the existing
workstation’s configuration, people can find the unused IP addresses and
required IP router address, then hook up their laptop, for example.
When a DHCP server is up and running, it provides a convenient environment
for mobile users. Users can hook up their laptops and instantly get the access
to corporate network. In general, DHCP users are anonymous to the
administrator. In a company where the basic TCP/IP security guard is
required, the basic DHCP environment is not good.
That is probably the one reason why large customers are not eager to use
DHCP servers. To breakthrough this problem, we must have another
mechanism in addition to DHCP. Presecured DDNS client and domain design
© Copyright IBM Corp. 1998, 2000
347
is the answer to this security problem. The TCP/IP administrator might have
slightly more workload than the static IP, but the result would be a well
protected dynamic IP network that is more secure.
8.2 RSA public key authentication system
Dynamic DNS, as defined in RFC 2137, uses an RSA public key/private key
authentication system to secure the dynamic DNS update. A client has a set
of keys stored on the hard disk. The key pair is dynamically generated by the
client program, NSUPDATE.EXE, or is generated by the administrator through
the DDNS Administrator GUI called DDNS Server Administrator. Figure 251
illustrates how the DNS dynamic update is protected with the digital signature
authentication.
Figure 251. RSA public key authentication system
A client owns the hostname so it authenticates the A RR (Address Resource
Record) in the DNS domain file. If the dynamic zone is created as a Dynamic
Secured one, clients can dynamically register A RR at any time. That means
the client generates a RSA key pair once at startup time and sends a
registration request to a DNS server together with its public key (encoded)
and its IP address. The important matter is that the client keeps the private
key on a hard disk or a networked drive and it should not be exposed.
348
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The RSA key system depends on the fact that the digital signature can be
verified with the client’s public key, and the private key cannot be discovered
from either the public key, or the digital signature, or a combination of both.
The client must generate a new digital signature each time an update request
is sent to DNS.
Once a hostname is registered with the public key and the digital signature,
the entry is verified or authenticated at each DHCP renewal time or when the
DNS TTL (Time-to-Live) is expired.
If the dynamic zone is defined as a Dynamic Secured one, clients cannot add
or update their hostname dynamically. Since the administrator defines
hostnames and generates key pairs for each hostname, the client should
import the specific key pair and must use it for further processing.
Figure 251 on page 348 simplifies the record entries of both domain file
(forward mapping file) and reverse file (reverse mapping file), but, basically,
there are two SIG records and one SIG record that authenticates the KEY
record. Another SIG record authenticates either the A record or PTR record.
The part of actual domain file is shown in Figure 252.
client1
4660
4660
4660
4660
3600
4660
IN HINFO "Ni0weDAwMDBlNTY4N2Y2MA==" "IBMDDNS-PROXY"
IN SIG A 1 4 4660 2147483647 893107703 0x6d25
client1.austin.cooking.net u1SD2P.....G5VK47mig== 2147483647
IN SIG HINFO 1 4 4660 2147483647 893107703 0x6d25
client1.austin.cooking.net HrwTz........e4p+a/g== 2147483647
IN SIG KEY 1 4 4660 2147483647 893107703 0x6d25
client1.austin.cooking.net FFqsEa.......Yujelkw== 2147483647
IN KEY 0x0000
0
1 AQPqq+gV....VuGWzJW3X
IN A 192.168.7.22
;Cl=3
;Cl=3
;Cl=3
;Cl=3
;Cl=3
;Cl=3
Figure 252. [Warp Server] Example of KEY, SIG and a resource record
The highlighted field in a KEY record is an encoded public key of the client.
Due to space limitations, the whole key is not shown here. The actual key
length is 1024 bits (128 bytes), but it is encoded to 88 bytes. A digital
signature is also 1024 bits but again encoded to 88 bytes.
The U.S. Government approved the 1024-bit RSA key system for export,
since it is only used for authentication and not to encrypt messages. For the
purpose of message encryption, the key length is restricted to 56 bits for the
export version and 128 bits for the U.S. domestic version only.
Chapter 8. Security of DHCP and Dynamic DNS
349
We distinguish two different dynamic types of DNS zones:
• Dynamic secured
The dynamic secured zone provides a great deal for a zone. Once a
client’s hostname is dynamically registered in the DDNS database, it is
protected and cannot be taken over by another client. The workstation with
the DDNS client function can dynamically join the network and can
register its name. Because of that, the DNS zone is open for all clients.
The dynamic secured zone buys tremendous flexibility (given the goal of
an administratorless system) in that it requires no administrator
intervention to have the DHCP clients registered, which is exactly the point
of DHCP. The only shortcoming of this concept is that a malicious client
could create useless names, however, this shouldn’t have an undesirable
effect on the network.
Be aware of the fact that any DDNS client could perform the very first
registration and then, suddenly, would own a hostname that might
otherwise be used for other server names.
• Dynamic presecured
If you are looking for the perfect secured DNS zone you need to look at a
so-called dynamic presecured zone. Clients joining the presecured
domain must have the key pair generated by the administrator. This key
then can be distributed to the client through a networked drive, through
e-mail, or through a diskette. Using this method, there is no chance for
clients to [unlawfully] join the DNS space. The administrator would have
perfect control over their DNS zone.
As for the client’s public key, it is not a secret one. It is intended to be known
by everyone. For example, the client’s public key can be retrieved through the
NSLOOKUP interface. Figure 253 on page 351 shows an example of the NSLOOKUP
output that provides all the information of client1.austin.cooking.net.
350
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Default Server: fajita.austin.cooking.net
Address: 192.168.7.10
> > Server: fajita.austin.cooking.net
Address: 192.168.7.10
client1.austin.cooking.netCPU = Ni0weDAwMDBlNTY4N2Y2MA==OS = IBMDDNS-PROXY
client1.austin.cooking.net
Signature Record covering A RR's
Authentication Algorithm = 1 (MD5/RSA) Labels = 4
Original TTL
= 4660 (1 hour 17 mins 40 secs)
Client SIG expiration = 2147483647, Mon Jan 18 21:14:07 2038
Time signed
= 893107703, Mon Apr 20 16:28:23 1998
Server SIG expiration = 2147483647, Mon Jan 18 21:14:07 2038
Key footprint = 0x6d25
Signer's name = client1.austin.cooking.net
Signature = u1SD2P0kC8Idl0HyelcqqZIRrMEkHQ3e2SdmwaaStBRRcSBC4DDRMSIRDwvgwcrA7wfDvY9
QQEoOKG5VK47mig==
client1.austin.cooking.net
Signature Record covering HINFO RR's
Authentication Algorithm = 1 (MD5/RSA) Labels = 4
Original TTL
= 4660 (1 hour 17 mins 40 secs)
Client SIG expiration = 2147483647, Mon Jan 18 21:14:07 2038
Time signed
= 893107703, Mon Apr 20 16:28:23 1998
Server SIG expiration = 2147483647, Mon Jan 18 21:14:07 2038
Key footprint = 0x6d25
Signer's name = client1.austin.cooking.net
Signature = HrwTzfLIGkvFfUmHwKr6aaHV/qO6RVc6uzItH9cYYQs1d1PRVgGrENlQZscj2gwCY4Z
B/LfBcH1fCVse4p+a/g==
client1.austin.cooking.net
Signature Record covering KEY RR's
Authentication Algorithm = 1 (MD5/RSA) Labels = 4
Original TTL
= 4660 (1 hour 17 mins 40 secs)
Client SIG expiration = 2147483647, Mon Jan 18 21:14:07 2038
Time signed
= 893107703, Mon Apr 20 16:28:23 1998
Server SIG expiration = 2147483647, Mon Jan 18 21:14:07 2038
Key footprint = 0x6d25
Signer's name = client1.austin.cooking.net
Signature = FFqsEaYujelrDfRxlS77NxE5Om/l+OgvRUN/P2QLPiNIYa5S1FdNBttBi3lOQRRs6eRje
fwjJFBlIqqPVuX8kw==
client1.austin.cooking.net
flags = 0x0000 (HostKey) protocol = 0, algorithm = 1 (MD5/RSA)
public-key data = AQPqq+gVMx4td+vglqCjv3fknqVyNn++w/uAoNHA/t52qzVkyXOYAGftLS78gi
pWJtVpDGiqMWs7ly3VuGWzJW3X
client1.austin.cooking.netinternet address = 192.168.7.22
austin.cooking.netnameserver = fajita.austin.cooking.net
fajita.austin.cooking.netinternet address = 192.168.7.10
>
Figure 253. [Warp Server] NSLOOKUP output returning public key
8.3 Getting more information from the client to DNS
With TCP/IP Version 4.1 for OS/2, the OS/2 DDNS client has a new function
that can force the client to input more practical information to the DNS
database:
Chapter 8. Security of DHCP and Dynamic DNS
351
• option 192
Option 192 is defined in the DHCP server configuration file, and it enables the
DHCP client to load a special window for the user to input predefined data,
such as:
• Office and building number
• Name
• Phone number
• E-mail address
Figure 254 on page 352 shows how the administrator defines option 192 at
the DHCP server.
Figure 254. [Warp Server] DHCP server, configuring option 192 txt record
To enable option 192, type the required input data in the text string containing
tablets field. For example:
Office;Name;Phone;Email-ID;
352
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
When the DHCP client gets control from the server, the server initiates the
DDNS client to prompt the user to input data, as shown in Figure 255 on page
353.
Figure 255. [Warp Server] DDNS client prompts for input
8.4 Presecured domain
The presecured domain is the most secured DNS name space available. For
information on how to set up a presecured domain, refer to 4.2.6.3, “Dynamic
presecured mode” on page 168. You can set up a subdomain in the corporate
root domain and make the sub-domain work as a dynamic presecured
domain using the TCP/IP 4.1 DDNS.
Two reasons for a higher security standard would be:
• The DHCP client cannot register its hostname without having a RSA key
pair and the DDNS client code present.
• Without a RSA key pair, the PTR record for reverse mapping cannot be
registered either, so that reverse mapping would fail, and the administrator
would be able to know which IP addresses are unofficial users. The
administrator would be able to determine the MAC address of those DHCP
clients’ LAN adapters. Then, at another step, the administrator could
configure the DHCP server to exclude those unlawful clients through the
server configuration GUI.
Just as a Lotus Notes administrator generates an ID file for the user and
distributes it to users, the TCP/IP administrator generates a key file and
passes it to the client.
Chapter 8. Security of DHCP and Dynamic DNS
353
Figure 256. Presecured TCP/IP domain concept
Although the TCP/IP administrator would have a higher workload establishing
a presecured domain in comparison to working with a static DNS
environment, the presecured domain ensures a well-trusted DNS name
space, and it can reduce the total cost of administration.
The most reasonable implementation of a presecured domain is to combine it
with static DNS servers and create a kind of hybrid domain design, as shown
in Figure 257 on page 355.
354
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 257. Combination of static DNS and dynamic presecured DNS
This design ensures that an administrator pre-generates hostnames and
corresponding key files in advance. Then, the administrator would create a
CNAME (alias) record in the root DNS table when a new request for the
hostname is received.
The following is the example of CNAME entries in the root DNS table:
uwe-zim
michaelmd
IN
IN
CNAME
CNAME
usr001.armonk.cooking.net
usr002.austin.cooking.net
Let’s look at the hostnames from a generic view. The user uwe-zim would
always be known in the network as uwe-zim, regardless of what the hostname
in the FQDN looks like. You can set up the all DHCP/DDNS servers in your
network using this setup scheme. When the client machine is moved to
another location, the location’s DHCP server will serve the new IP address for
it, but the client machine can still update its home DDNS server with the new
IP address.
This design allows the safe and smooth deployment of Dynamic DNS
together with DHCP servers.
Chapter 8. Security of DHCP and Dynamic DNS
355
8.5 ProxyArec consideration
The ProxyArec function is desired by the industry, and several companies
provide products that embrace ProxyArec. ProxyArec is based on the IETF
Internet draft called Interaction between DHCP and DNS. This draft can be
and found on the Internet at the following Web site:
http://www.ietf.org/internet-drafts/draft-ietf-dhc-dhcp-dns-08.txt
The latest draft may have a newer number.
Besides the TCP/IP Version 4.1 DHCP and DDNS servers, we also tested
NTS Shadow IPserver, which also provides a ProxyArec function. However,
NTS Shadow IPserver does not support option 81, which tells the DHCP
server that a DHCP/DDNS client wants to register/update its A RR by itself;
so the DHCP server would not perform the ProxyArec function to update the
DDNS table. Option 81 is discussed in 8.5.1 on page 358.
When the DHCP server is configured to enable the A resource record update
in addition to the PTR resource record update, sometimes called a ProxyArec
function, any DHCP client which carries the option 12 hostname will register
both A RR and PTR RR. Since the DHCP server uses its own RSA key for all
DDNS update requests, there is no authentication mechanism to validate that
the client really is the client that has the original hostname.
However, the ProxyArec function in TCP/IP Version 4.1 for OS/2 DHCP server
provides security. When the DHCPSD.CFG file in the \MPTN\ETC
subdirectory has the following statement, the client’s MAC address will be
used to memorize the ownership of the A RR:
ProxyArec Protected
To disable the MAC address identification, the statement should be:
ProxyArec Standard
Then, the DHCP server will blindly update the A RR without comparing the
MAC addresses at all.
From the DHCP Server Configuration GUI, the ProxyArec settings require you
to complete two panels. First, you need to specify a domain name and a
DDNS server IP address (or hostname) in the DHCP Server Parameters
window, as shown in Figure 258 on page 357.
356
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 258. [Warp Server] ProxyArec - Enable DHCP server to update A RR
The Global Parameters window, as shown in Figure 259, is the only place
where you would enable ProxyArec. You must have a dynamic DNS server to
enable dynamic A RR updates.
Figure 259. [Warp Server] Enable ProxyArec and option for protection
Chapter 8. Security of DHCP and Dynamic DNS
357
The question in Figure 259 on page 357 that says Automatically update DDNS
A record on client’s behalf if requested? is defaulted to No. If you select Yes
from the pull-down menu, the ProxyArec function is enabled. Another
selection box (Verify client ID before performing update?) decides on the two
modes of ProxyArec:
• Standard
• Protected
To enable protected ProxyArec, select Yes from the pull-down menu.
Remember that the main purpose of the ProxyArec function is to support
non-DDNS clients to dynamically register/update its hostnames where
security is not a first priority. Configuring the DHCP server to use the client’s
MAC address to protect unauthorized updates from a different client does not
indicate security. Dynamic DNS name space is almost like a public space.
Any DHCP client can register the hostname with the DHCP option 12.
Windows 95 and NT handle option 12 in a different way than OS/2. Even
though the Windows 95/NT client was provided with a hostname in the
TCP/IP DNS settings, the hostname is not used. Instead, the Computername
definition is used as option 12. This is a unique (proprietary) Microsoft
implementation of a hostname. Traditionally, the computer name, also known
as the NetBIOS name, is not controlled by the IS department or TCP/IP
administrators; so, there might be a chance that people use the same
computername in a TCP/IP network, which usually results in duplicated
names. Duplicate names result in not opening LAN adapters.
For your information, WorkSpace On-Demand Release 1 uses the ProxyArec
function when DDNS is required. IBM PPP server on top of Warp Server’s
Remote Access Server also uses the ProxyArec, because it supports
non-LAN distance clients, such as Windows 95 or Windows NT Workstation.
8.5.1 ProxyArec and option 81
We don’t recommend having a ProxyArec DHCP server and standard DDNS
clients in a same location because a DDNS client cannot use its own interface
to the DDNS server. An RR will be owned by the DHCP server. However, in
such a situation, the OS/2 DHCP/DDNS client can use a special DHCP
option, which is option 81, which would disable the ProxyArec function for a
requesting client.
Unfortunately, there is no graphical user interface where you could define
option 81. You must manually add option 81 to the
358
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
\MPTN\ETC\DHCPCD.CFG file using HEX translated FQDN (Fully Qualified
Domain Name). This is shown in Figure 260.
# Basic options required
clientid
MAC
interface lan0
# Uncomment as desired for logging
numLogFiles
4
logFileSize
100
logFileName
dhcpcd.log
logItem
SYSERR
logItem
OBJERR
logItem
PROTERR
logItem
WARNING
option 12 tuna
# Host name
option 81 HEX "00 00 00 74 75 6E 61 2E 61 75 73 74 69 6E 2E 63 6F 6F 6B 69 6E 67 2E 6E 65 74"
# Don't ProxyArec tuna.austin.cooking.net
updateDNSA "nsupdate -h%s -d%s -s"d;a;*;a;a;%s;s;%s;3110400;q" -q"
updateDNSTxt "nsupdate.exe -h%s -s"d;txt;%s;a;txt;%s;s;%s;3110400;q""
Figure 260. [OS/2 Warp] DHCPCD.CFG Example with option 81
We would prefer to specify an ASCII name rather than HEX code. But it would
not work. The IETF document Interaction between DHCP and DNS defines
option 81, as shown in Figure 261:
Code
81
Length Flags Rcode1 Rcode2
n
00
00
FQDN
00
Figure 261. IETF definition of option 81
The flags indicates the DHCP server to do the following:
00
01
Client wants to be responsible for updating the FQDN to IP
address mapping
Client wants the server to be responsible for updating the FQDN to
IP address mapping
Rcode1 and Rcode2 define the response code area used by the DHCP server
to check a DHCP client’s response for Dynamic DNS updates.
Chapter 8. Security of DHCP and Dynamic DNS
359
8.6 Securing lease allocations
There is no one solution for controlling which clients can obtain a DHCP lease
from a server, or if they should get a lease at all. However, there are some
alternatives available to you:
• You can use classing (option 77) when configuring your address pools.
Find an example on classing in Figure 237 on page 326.
Using classes requires you to visit each client and modify their
configuration. Once you have your classes established, only devices that
belong to the class will be allocated a lease.
Be aware that not all clients can make use of option 77.
• At the server, you could allocate IP addresses to individual MAC
addresses. This can be very labor intensive, but ensures that only
machines with known MAC addresses will obtain a lease from a server.
• You may be able to find DHCP server software that allows you to list which
MAC addresses the server will accept. DHCP servers that support
roaming machines may be adapted to such use.
• If you want to ensure that clients are only served by a particular server,
you could conceivably install one DHCP server per subnet. This is not a
good solution, but if you choose to do this, ensure that your routers do not
forward broadcast traffic, and that there is no DHCP Relay Agent on the
subnet.
This solution does not stop any malicious user from connecting to the
network and obtaining an IP address.
8.6.1 Preventing access to unauthorized devices
This would have to be done using a mechanism other than DHCP. DHCP
does not prevent other clients from using the addresses it is set to hand out
nor can it distinguish between a computer's permanent MAC address and one
set by the computer's user. DHCP can impose no restrictions on what IP
address can use a particular port nor can it control the IP address used by
any client. You may be able to use classing to limit leases to only those clients
that belong to a valid class.
8.6.2 Rogue DHCP servers
It is possible that a malicious or inexperienced user could create problems on
your network by setting up an unofficial DHCP server.
360
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The initial problem is that the server could pass out IP addresses already
belonging to some other computer. It is possible that you could end up with
two or more devices ending up with the same IP address. The end result is
problems using the devices. These problems can be intermittent or the
devices may fail entirely, necessitating a restart.
Other problems are possible if the unofficial server gets a device to accept its
lease offering, and then passes to the device its own (possibly incorrect)
DHCP options. For example, if options 1 or 3 are incorrect, the device may
not be able to communicate with devices outside its own subnet (or in some
cases, with devices on the same subnet)
Another scenario to consider is a device that loads its operating system over
the network using TFTP. If this device is directed to load a different file
(possibly on a different server), it allows a perpetrator to take over the client.
Given that boot parameters are often made to control many different things
about the computers' operation and communication, many other scenarios
are just as serious.
Note that by using BOOTP you are exposed to the same vulnerabilities.
8.6.3 Connecting to Untrusted Networks - Firewalls
There may come a time when you want to connect your private intranet to an
untrusted network. What is an untrusted network? The biggest example is the
global public Internet (maybe it is time to have your own local mail or Web
server rather than having your ISP host it for you). Another example may be
one of your customers’ networks. Say you are a consulting firm who wants to
provide remote IS services for your clients. One way to do this is to connect
your two corporate networks together. You would not want all your customers,
and whomever they have logged onto their networks to have access to all
machines on your private network. You probably want them to have access to
your Web server and not much else. However, you need access to machines
on their network so that you can do the appropriate maintenance, upgrades,
monitoring, and so forth, to perform your IS duties. There are many more
examples of why you might want this kind of protected architecture between
your private intranet and other untrusted networks. This section will show you
the how-to more than the why do it. Be aware that there are more methods to
connect than the single one we are showing you.
To interconnect safely, you will need a firewall. A firewall is not a panacea for
security, and you should not treat it as such. A firewall can help provide
security, but there are many more issues beyond the firewall to keep in mind.
Chapter 8. Security of DHCP and Dynamic DNS
361
Firewalls provide secure access from a company’s internal TCP/IP network to
the public Internet. Intruders are blocked and cannot access the internal
network from the public network without authorization. Internal users,
however, can still access resources on the public Internet. In combination with
good security practices, firewalls can help keep private information on your
network secure.
Several vendors produce high quality, full-featured firewalls. Examples
include Checkpoint Firewall-1 for AIX, NT, HP-UX, and Solaris, IBM Firewall
for AIX and Windows NT, and LanOptics Guardian for OS/2 Warp and
Windows NT. Alternatively, many firms opt to contract for firewall services with
their network providers to avoid having to learn the specialized skills required
to manage network security.
Most firewalls can support two proxy methods to mask internal client IP
addresses from the Internet. SOCKS is the most modern method, and it can
provide transparent access for all TCP/IP-based applications if the client
operating system has a SOCKSified protocol stack. OS/2 Warp and
WorkSpace On-Demand have built-in SOCKS support; third parties provide
SOCKS extensions for Windows 95 and Windows NT clients. Practically all
Web browsers support SOCKS-based firewall access.
Proxy servers can also provide access to public sites through a firewall for
internal clients. However, proxy servers can only handle HTTP, FTP, and
Gopher protocols, so they are mainly of benefit to Web browsers. Caching
proxy servers, such as the IBM Web Traffic Express, can even cut down on
network traffic through the firewall by keeping frequently accessed Web
pages and files on the proxy server itself. Web Traffic Express is now part of
the IBM Websphere Performance Pack.
In principle, a proxy server, such as IBM Web Traffic Express or Domino Go
Webserver, can be used as a simple firewall solution when some degree of
security is needed. A proxy server performing that role on Warp Server, for
example, would need two network adapters (one for the internal network and
one for the external network) and a proxy server software package installed
and running. That server should have an IPGATE OFF command in its
\MPTN\BIN\SETUP.CMD file so that TCP/IP traffic will not be forwarded from
one LAN adapter to the other. (Otherwise, an intruder could access internal
systems.) Also, it should not have any daemons running which may expose
vulnerabilities. Nor should the proxy server have any LAN services bound to
the adapter handling external network access. Moreover, the proxy server
should not handle any requests from external users. Provided these
conditions can be met, a proxy server can act as a simple firewall. Note,
however, that this solution will only keep client identity safe by retrieving
362
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
information from the Web by proxy, and will not perform any of the advanced
firewall functions to allow you to scrutinize suspect information packets or
keep logs of all traffic passing to and from the Internet.
You may wish to download a 60-day evaluation version of IBM Web Traffic
Express to learn how to set up a caching proxy server. Please visit:
http://www.software.ibm.com/webservers/wte/index.htm
to obtain your own evaluation copy.
Naturally, commercial grade firewall products go well beyond simple proxy
server functions. For a full description of how to set up a firewall and institute
good security management, we recommend consulting a redbook specifically
devoted to firewall security, such as Check Point FireWall-1 on AIX A
Cookbook for Stand-Alone and High Availability, SG24-5492 or A
Comprehensive Guide to Virtual Private Networks, Vol.1: IBM Firewall, Server
and Client Solutions, SG24-5201.
8.6.4 Connecting through untrusted networks - VPN
A virtual private network (VPN) utilizes IP tunneling to create an extension of
an enterprise's private intranet across a public network such as the Internet.
VPNs securely convey information across the Internet connecting remote
users, branch offices, and business partners into an extended corporate
network.
Internet Service Providers (ISPs) offer cost-effective access to the Internet
(via direct lines or local telephone numbers), enabling companies to eliminate
their current, expensive leased lines, long-distance calls, and toll-free
telephone numbers.
A 1997 VPN Research Report, by Infonetics Research, Inc., estimates
savings from 20% to 47% of wide area network (WAN) costs by replacing
leased lines to remote sites with VPNs. And, for remote access VPNs,
savings can be 60% to 80% of corporate remote access dial-up costs.
Additionally, Internet access is available worldwide where other connectivity
alternatives may not be available.
The technology to implement these virtual private networks, however, is just
becoming standardized. Some networking vendors today are offering
non-standards-based VPN solutions that make it difficult for a company to
incorporate all its employees and/or business partners/suppliers into an
extended corporate network. However, VPN solutions based on Internet
Chapter 8. Security of DHCP and Dynamic DNS
363
Engineering Task Force (IETF) standards will provide support for the full
range of VPN scenarios with more interoperability and expansion capabilities.
The key to maximizing the value of a VPN is the ability for companies to
evolve their VPNs as their business needs change and to easily upgrade to
future TCP/IP technology. Vendors who support a broad range of hardware
and software VPN products provide the flexibility to meet these requirements.
VPN solutions today run mainly in the IPv4 environment, but it is important
that they have the capability of being upgraded to IPv6 to remain
interoperable with your business partner's and/or supplier's VPN solutions.
Perhaps equally critical is the ability to work with a vendor who understands
the issues of deploying a VPN. The implementation of a successful VPN
involves more than technology. The vendor's networking experience plays
heavily into this equation.
For a more in depth discussion of VPN theory and implementation see the
redbooks A Comprehensive Guide to Virtual Private Networks, Volume II: IBM
Nways Router Solutions, SG24-5234 or A Comprehensive Guide to Virtual
Private Networks, Volume III: IBM Cross-Platform and Key Management
Solutions, SG24-5309.
8.6.5 TFTP security
If you decide to provide BOOTP services, remember that TFTP is inherently
insecure. In TCP/IP Version 4.1 for OS/2, you can specify which client IP
addresses have access to the TFTP directory. For example, our
\MPTN\ETC\TFTPAUTH file contained the following lines:
C:\TFTPBOOT RO 192.168.6.17
C:\TMP RO 192.168.6.11
which means that the machines with the respective IP addresses have read
only access to the directories listed. Since you can preassign IP addresses
based on the machines MAC address, this is fairly safe.
In TCP/IP Version 4.21 for OS/2 you can specify either the client hostname or
IP address when defining access for the TFTP directory. Clients may have
Read-Only Access or Read Write Access depending on the requirements.
Please see Figure 58 on page 84, where we configured TFTP access for a
specific IP address. In “BOOTP” on page 79, we describe how to configure
the TFTP Daemon and its security features.
364
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 9. Reliability
"I can’t get on the network!"
Those six words should never be heard by a network manager. Yet with
TCP/IP networks, they are, unfortunately, all too common.
This chapter focuses on how to make your TCP/IP network as reliable as
possible, with connections readily available and conflicts eliminated. In
particular, we'll explain how to create fault tolerant TCP/IP address servers.
If your TCP/IP network does not need to be reliable, you can skip this chapter.
9.1 Battlefield questions
To design a network for reliability, you might think of your network as a
battlefield. The goal is to assure that your messenger can travel from one end
of the battlefield to another without getting blown up. In between are roads,
bridges, highways, airways, rail lines, and other mechanisms used by your
soldiers to communicate. If one (or more) of these conveyances gets blown
up, your courier must still reach the destination.
In designing your network, you should ask yourself a series of What if?
questions. What if the router shorts out? What if the DHCP server is buried by
a mudslide? What if a remote office satellite link gets blocked because of
sunspots?
Then, when you have examined the What if? questions, there are some more
questions to answer:
• How likely are these events?
• How much reliability do I need? For whom?
• Where can I spend limited dollars to get the highest possible degree of
reliability?
• If a failure does occur, will I know about it? How fast can I respond?
• Do I regularly rehearse how to respond to failures?
• Do I have a written plan of action to handle network outages?
• Am I measuring the frequency and severity of network problems so I can
determine whether service improves?
© Copyright IBM Corp. 1998, 2000
365
If you’re doing your job, these questions never have permanent answers. You
should always regularly reexamine these issues, even if your network runs
smoothly.
9.2 Failure events
Datagrams can only travel across your network as long as every part of the
link works, from end-to-end. Your network can only be as reliable as the
weakest link.
9.2.1 Severed connections
One obvious possible failure is a severed connection. Wires may be cut,
cables inadvertently unplugged, or fiber optics dug up by a backhoe. IP can
easily reroute around such failures, provided another connection is available.
An easy way to design redundancy into connections is to simply provide two
(or more) separate lines or wires, preferably using different technologies.
Another way is to structure the network as one unbroken ring (or circle). If
one connection in the ring does break, traffic can still flow throughout the
network because all points are still connected.
The U.S. Federal Aviation Administration, for example, discovered that their
connection redundancy failed, and air traffic control services were disrupted
for a time. (Fortunately, no one suffered any injury.) The FAA uses telephone
lines to connect various facilities. These facilities each have multiple circuits,
and switching equipment can quickly reroute network traffic over backup
circuits. However, the FAA granted a single long distance telephone company
the contract to connect their facilities. When the telephone company’s own
network failed, the FAA had no backup, and the multiple circuits didn’t matter.
Now, the FAA has contracted with a second long distance telephone company
to provide backup connections should they be needed.
In another example, PanAmSat’s Galaxy IV satellite suffered a catastrophic
failure in 1998. Amazingly, almost none of the U.S. paging services had any
immediate backup satellite, so paging services were disrupted for days.
Physicians, in particular, were hard hit, because they lost contact with
hospitals and could not be called in by pager.
9.2.2 Facility loss
Fires, floods, earthquakes, hurricanes, power outages, and other
catastrophes can wreak havoc with a network by knocking out crucial servers,
routers, bridges, switches, and other devices. Large parts of the Internet, for
366
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
example, were disabled when the last San Francisco earthquake knocked
down some key buildings.
Many companies do not properly plan for these contingencies. Your network
may be required to handle these events, so try to avoid placing all your vital
network devices in one location.
An uninterruptable power supply (UPS), with a remote alerting mechanism,
can help keep critical network systems up-and-running. However, backup
electrical power should be provided for all critical systems. It’s easy to forget
that routers, bridges, concentrators, hubs, servers, switches, and even clients
all need backup power if they are to continue operating.
9.2.3 Router outages
Subnets in a TCP/IP network can quickly become cut off from the rest of a
network if a router fails. These devices should be extremely reliable, and
many network designers choose dedicated pieces of industrial equipment
from vendors such as Cisco (or others) instead of less reliable PCs. Most
dedicated routers use special operating systems that help to facilitate their
speed and reliability. Routers can also be installed in pairs.
9.2.4 DHCP server problems
DHCP can help provide added flexibility and convenience in managing a
TCP/IP network. Static addresses can often conflict, locking systems out of a
network. Also, static configurations can only be changed with a significant
amount of work, making it difficult to deal with router outages, changes in
subnet structures, and so forth. In other words, DHCP servers can actually
help ensure reliability.
However, if a DHCP server fails, then a TCP/IP network can quickly run into
trouble. DHCP servers must be reliable. (See 9.3, “Redundancy and fault
tolerance in DHCP” on page 369, for one way to ensure DHCP server
reliability.)
Failures of DHCP relay agents can also prevent new systems from joining a
TCP/IP network. We recommend using DHCP relay agents sparingly. Instead,
you may wish to have dedicated routers handle the task with BOOTP relays
as needed. If you do use DHCP relay agents, consider a more fault tolerant
solution, such as the one described later in this chapter.
Chapter 9. Reliability
367
9.2.5 Name server difficulties
Without name servers, your users will have to resort to numeric IP addresses,
effectively making your network useful only to the most savvy technicians.
DNS servers must also be active and reachable at all times.
TCP/IP clients, including OS/2 Warp 4 and Windows 95, can readily accept
two or more name server addresses. If the first name server can’t be
reached, the client will attempt to contact the backup name server(s). We
recommend taking advantage of this simple feature and to use two (or more)
name servers on your mission critical network.
Dynamic DNS servers, such as OS/2 Warp Server, also help minimize
network problems. For example, if an important Web server fails, its name
can be reassigned to another numeric IP address that much more quickly and
easily with a DDNS server. In fact, this secure update may be triggered
automatically when a server failure has been detected by network alerting
tools.
9.2.6 Other server vulnerabilities
Server failures can prevent users from accessing important applications and
databases. Technologies such as RAID can help cope with hard disk failures.
Many vendors provide fault tolerant server solutions, including IBM. Storage
management software, such as the IBM ADSM, help send backup copies of
files across the network to off-site locations.
Yet with all these server technologies, it’s important to remember that most
failures occur because of software problems. All the RAID storage in the
world won’t help if your server suffers from the dreaded blue screen. Some
operating systems can be vulnerable to failure, especially when intentionally
induced. For example, in the late 1980s, Robert Morris at Cornell University
discovered a security vulnerability in some versions of UNIX. He used this
vulnerability to spread a worm which quickly infected numerous systems
connected to the Internet. For several hours, the Internet essentially ground
to a halt as system managers attempted to address the problem. Many could
not even log onto their own servers since the worm multiplied rapidly,
consuming practically all the attention of the server’s processor(s). More
recently, some versions of Windows NT have been vulnerable to attack using
ping and/or Telnet.
You should keep close watch on your servers and regularly review software
bug reports and other information in case others find additional problems
which may affect you. Antivirus software, such as Norton Antivirus, may
368
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
detect and eradicate PC viruses on your servers. (Viruses can also cause
network problems.)
9.2.7 Client failures
Many network problems really relate to the fragility of traditional PC clients,
which tend to break down all too often. It isn’t hard to delete an important
icon, modify a critical system file, or otherwise render a traditional PC unable
to connect to the network. PC users then have to wait for time consuming
repairs to software, hardware, or both.
Various studies estimate the cost of managing a typical business PC (the
annual total cost of ownership) at between $5,000 and $26,000. These costs
appear to be rising despite (or perhaps because of) advances in technology.
While that annual cost may be perfectly justified for many users, including
programmers, engineers, heavy office suite users, hobbyists, and others, the
high cost is not necessarily attractive to everyone. Yet, the benefits of PC
applications, graphical user interfaces, and online help do make sense.
Many companies in the computing industry recognize these problems and are
working to solve them through network computing. True network computing
solutions provide most or all of the benefits of PC applications with a total
cost of ownership approaching that of a mainframe or UNIX terminal.
Examples include network computers from IBM, Sun, Oracle, and others,
Java, and the IBM WorkSpace On-Demand for Intel-compatible systems. One
of the key benefits is that users can log on from any station on the network
and get access to individual documents, applications, and Web pages from
anywhere. Consequently, network computing can improve the reliability of the
whole network by providing access to it more often and with less complexity.
While some users do need the mainframe on a desk that the PC has become,
we recommend seriously examining network computing solutions for many
users. We also recommend implementing network computing solutions, even
for PC users where they make sense, such as Web-delivered Java
applications, to help promote reliability within your network.
9.3 Redundancy and fault tolerance in DHCP
DHCP servers must be reliable if the network is to be consistently available.
As leases expire or users reboot their workstations, the impact from a DHCP
server outage can spread quickly. Eventually the entire network can grind to a
halt. Naturally, you should choose a reliable, robust and sturdy platform to
host your DHCP services. Thoughtful configuration and consistent
administration of DHCP services in your network are equally important.
Chapter 9. Reliability
369
RFC 2132 is the most current specification for DHCP function. However, it
does not describe a means by which multiple DHCP servers can
cooperatively administer addresses. With network access depending on the
DHCP server, there is demand for this type of functionality. There are
currently several software products on the market that use proprietary means
for cooperative and redundant DHCP servers.
Additionally, the IETF’s dynamic host configuration work group has been quite
active recently. They are working to define a failover specification that would
also achieve redundancy in DHCP servers. The latest draft available at this
writing is draft-ietf-dhc-failover-05.txt.
9.3.1 The challenge of fault tolerance
DHCP fault tolerance presents some challenges for both the currently
available proprietary methods, as well as the proposed RFC standards.
Fault tolerant DHCP services can be provided in one of two ways:
Cooperative redundancy involves multiple servers that assign resources from
the same pool, concurrently or in a primary/standby configuration.
Non-cooperative redundancy uses multiple servers and multiple
non-overlapping pools.
If two or more servers are to administer address from the same pool it is
important that each is constantly aware of the other’s activities. So to ensure
the reliability of their interconnection, one might place the servers very close
together in the network. However, this arrangement is in direct conflict with
another goal of redundancy, which encourages us to put some distance
between redundant resources. With redundant servers far apart, we achieve
better availability to the clients if a network problem occurs.
9.3.2 Cooperative redundancy
Consider the following scenario. A medium sized network has two DHCP
servers cooperatively assigning configuration to clients. A fault occurs that
results in a partitioning of the network. We now have two running, but
disconnected, networks. Fortunately, there is a DHCP server on each
network, so at first glance it may seem that our clients are minimally
impacted.
Now we have two groups of clients and DHCP servers, and a shared address
pool. Prior to the network partitioning, the servers were in constant contact.
Each was aware of the other’s address assignment activity. Now this
communication can not take place. Depending on how the code is written
within the DHCP servers, one of three things could happen next:
370
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• Because they can’t communicate with each other, both DHCP servers stop
working. All users are impacted.
• As a result of predetermined roles, one of the DHCP servers continues to
function while the other stops. A portion of the users are impacted.
• Both DHCP servers assume that its partner has failed and both continue
to assign addresses. No immediate impact to users occurs. But when the
network fault is repaired there is significant possibility of duplicate IP
addresses.
9.3.3 Non-cooperative redundancy
If you have plenty of available IP addresses for your network, you can easily
establish a redundant and fault tolerant environment using multiple but
independent DHCP servers. The simple solution here is to divide an address
pool between multiple servers. That is, each server has a portion of an
address pool that only it can assign. When one server fails, you lose only the
ability to assign the addresses it owns. As long as clients have connectivity to
another server, they can get a valid configuration.
This arrangement provides redundancy, yet does not require the DHCP
servers to have knowledge of each other’s activity. Because of this they can
be advantageously positioned in the network and continue to function during
connectivity faults. Of course, the downside to this method is the number of IP
addresses required.
The amount of extra addresses needed can be determined by dividing the
total number of available addresses by the number of DHCP servers. For
example, if you are assigning 100 IP addresses using four DHCP servers, 25
addresses should be reserved for redundancy. A maximum 75 addresses may
be active at one time.
9.3.4 Standby DHCP server arrangement
While it’s certainly possible to purchase packaged fault tolerant solutions,
such as Vinca’s StandbyServer for OS/2 Warp
( http://www.vinca.com/products/os2/os2_sbs.html) or an IBM AIX cluster (see
9.4, “AIX features” on page 380), we thought it would be helpful to show a
simple standby server solution using the features built into OS/2 Warp Server
for e-business and TCP/IP Version 4.21 for OS/2. By understanding the
basics, you can quickly graduate to more sophisticated solutions if needed.
9.3.4.1 Prerequisites
To implement a fault tolerant DHCP network you should have:
Chapter 9. Reliability
371
• Two systems running OS/2 Warp Server for e-business (with TCP/IP 4.21
and its DHCP server software installed). One of these systems, the
backup, may be an older, slower 486, as long as it has just enough
performance to keep the network up and running on an emergency basis
should the primary server fail.
• The backup DHCP server can, itself, be a DHCP client if it’s on the same
subnet as the primary. Otherwise, we assume your backup DHCP server
has a static IP address assigned on another subnet.
• With multiple subnets, your router(s) should include active BOOTP relay
agent(s), so that DHCP requests from clients will be handled by either the
primary or backup DHCP server as needed.
• For testing purposes, you may wish to have one DHCP client, of any type,
available on the network.
We also assume that your primary DHCP server is up and running correctly
and that Boot Manager is installed.
Although optional, we recommend also installing OS/2 Warp Server’s
Netfinity services (available through IBM Software Choice), at least on your
backup server, to process alerts so that you can be notified if the primary
DHCP server fails.
9.3.4.2 Step-by-step procedure
1. Enable REXECD and FTPD on the primary server. If REXECD is running
on the primary server, tasks can be started on the server from a remote
system. Likewise, if FTPD is running on the primary server, then files can
be transferred to/from the server across the network using FTP. Naturally,
you should password protect both of these services.
372
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 262. [Warp Server] Choosing Autostart Service for REXECD
The TCP/IP Configuration notebook controls whether REXECD and FTPD
will be automatically started and what user names and passwords they will
accept. Since the TCP/IP Configuration notebook is a Java application,
you can access the notebook at the server console or from any
Java-enabled system on the network. As shown in Figure 262, start by
clicking on the Autostart tab in the TCP/IP Configuration notebook. Then,
highlight REXECD and make sure the Autostart Service checkbox is
selected. Repeat the same procedure for FTPD. (Both can be started from
INETD; so, you shouldn’t have to click on any of the other radio buttons.)
Chapter 9. Reliability
373
Figure 263. [Warp Server] Security section in TCP/IP Configuration
Security settings for both FTPD and REXECD can be controlled by first
clicking on the Security tab (see Figure 263). Add or change users using
the list box on the left. In this example, we’ve defined a user name of
administrator with a password of password. Then, click on the buttons for
FTP Server and REXEC Server to enable access to these services for the
user administrator. (See Figure 264 on page 375.) At a minimum, you
should provide read-only access to the directory specified by the ETC
environment variable on the primary server, normally C:\MPTN\ETC.
374
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 264. [Warp Server] Configuring REXECD options
When you’ve finished making these changes, close the TCP/IP
Configuration notebook, save your changes, and restart the server.
When the server restarts, you may wish to verify that REXECD and FTPD
are operating properly by accessing the server using FTP and REXEC
commands at a client.
2. Copy FAILSAFE.CMD to the backup server and add to its Startup folder.
FAILSAFE.CMD runs on the backup server and continuously monitors the
primary. You should add a program object to the backup server’s Startup
folder to automatically run FAILSAFE.CMD each time the backup server
boots.
Figure 265 and Figure 266 on page 377 display the REXX program
FAILSAFE.CMD.
Chapter 9. Reliability
375
/* FAILSAFE.CMD
* by Timothy Sipples and Peter Degotardi
*
*/
/* FAILSAFE is designed to monitor a DHCP server and make sure that
* it is in continuous operation. If the DHCP server fails then the
* monitoring system takes over server responsibilities, shuts down
* the failing server, and sends an alert.
*/
/* The following three values are assumed valid for all needs: */
server_IP = "192.168.7.10"
user_ID = "administrator"
password = "password"
same_subnet = 1
check_every = 30
config_file = "C:\MPTN\ETC\DHCPSD.CFG"
address cmd
call RxFuncAdd "SysSleep","RexxUtil","SysSleep"
call RxFuncAdd "FtpLoadFuncs","rxFtp","FtpLoadFuncs"
call FtpLoadFuncs
say "Standing by to respond to any failure of server "server_IP"..."
do forever
rc = FtpSetUser(server_IP,user_ID,password)
rc = FtpGet(config_file,config_file,"Binary")
rc = FtpLogoff()
"@dadmin -h"server_IP" -u"password" -n > nul"
if rc \=0 then do
"@dadmin -h"server_IP" -u"password" -n > nul"
if rc \= 0 then leave
end
say date()" "time()": Check OK!"
call SysSleep check_every
end
Figure 265. [OS/2 Warp] FAILSAFE.CMD (Part 1 of 2)
376
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/* Server Failure Detected */
say date()" "time()": Houston, we have a problem..."
'@start /n rexec 'server_IP' -l'user_ID' -p'password' SETBOOT /T:NO
/B'
call SysSleep 5
if same_subnet then do
"@dhcpmon -t"
"@arp -f"
"@ifconfig lan0 "server_IP
end
'@del c:\mptn\etc\dhcps.ar'
'@del c:\mptn\etc\dhcps.cr'
'@start "DHCP Server" /n /min DHCPSD.EXE'
'@start "DHCP Server Administration" /min DHCPSCPS.CMD'
Figure 266. [OS/2 Warp] FAILSAFE.CMD (Part 2 of 2)
At the top of FAILSAFE.CMD, you can change critical values to match
those required on your particular network. These values include the
numeric IP address of the primary DHCP server, the user ID and
password for REXEC and FTP access to the server, whether the backup is
on the same subnet (1) or not (0), how often to interrogate the primary
server to make sure it’s still running, and the full path and file name for the
DHCP configuration file.
FAILSAFE.CMD spends most of its time in a do..forever loop. First, the
latest DHCP configuration file is copied to the backup server using REXX
FTP functions. Then, the backup attempts to run a simple DHCP
administrator report on the primary server. If the primary server doesn’t
answer, a second attempt is made. If both attempts fail, FAILSAFE.CMD
assumes the primary server has failed. If at least one attempt succeeds,
FAILSAFE.CMD goes to sleep for a number of seconds (the check_every
value), and the loop then repeats.
When a primary server failure occurs, FAILSAFE.CMD will perform several
tasks. First, an attempt will be made (using REXEC) to disable the primary
DHCP server by rebooting it with the SETBOOT command and placing it
indefinitely at the Boot Manager menu (/T:NO). Then, FAILSAFE.CMD
waits five seconds to give the primary server time to go offline. If the
backup server is on the same subnet as the primary, then the backup’s
DHCP address is relinquished (DHCPMON -t), the ARP table is flushed,
and the backup server assumes the primary server’s static IP address.
Chapter 9. Reliability
377
Finally, FAILSAFE.CMD deletes any old DHCP database files (DHCPS.AR
and DHCPS.CR) then starts both DHCPSD.EXE (the DHCP server) and
DHCPSCPS.CMD (the DHCP administration access task).
At some point in the primary server failure handling in this REXX script,
you should add one or more lines to generate an alert to make sure you
are notified by pager, e-mail, or autodialing modem that your primary
DHCP server has failed. Netfinity’s Alert Manager can help with this task.
Note that FAILSAFE.CMD generates on screen status messages
periodically. You can log these status messages by redirecting the screen
output of FAILSAFE.CMD to a file. For example, you could type in the
command:
FAILSAFE > DHCPBACK.LOG
to capture these messages in the file DHCPBACK.LOG.
Although this approach employs relatively simple TCP/IP technology, it
should help provide some added reliability to your network.
9.3.4.3 Technical caveats
• We have not included any mechanism for monitoring the backup DHCP
server to make sure it does not fail. Failure of the backup could leave your
network vulnerable, and you should be alerted should such a failure occur.
You may wish to use Netfinity services on the backup server as a
monitoring tool.
• We discovered that the backup server will not always start correctly if it
uses the full database from the primary DHCP server. Therefore, the full
database is not copied by FAILSAFE.CMD.
• Since we are not copying the full DHCP database from the primary server
to the backup, new clients attaching to the network when the backup has
taken over may receive an IP address assignment already in use by
another client. The client may have to reject several offers from the backup
DHCP server until an available address is found. This process can delay
final attachment to the network for new clients while the backup server is
in charge.
• FAILSAFE.CMD does not include some necessary error checking. For
example, you may wish to add error checking to the FTP file transfer. We
purposely kept FAILSAFE.CMD short in order to more clearly demonstrate
the basic concepts.
• A router or network connection failure, even a temporary one, could cause
the backup DHCP server to activate. Therefore, it’s possible that both the
primary and backup DHCP servers could be simultaneously active. (If the
378
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
primary DHCP server suddenly becomes unreachable, the SETBOOT
command will not be received by the primary, and it won’t shut down.) That
shouldn’t cause any immediate problem, particularly if the network
connection between the two subnets remains broken. Again, the goal is
simply to preserve reasonable network access until the primary failure can
be remedied at a more convenient time.
• We have not included a standby DDNS server example, because we
recommend that you simply provide each client (via DHCP) the IP
addresses of at least two name servers. Then, clients will automatically
use whichever name server is active. If for some reason you do need to
monitor your DDNS server(s), then you can use techniques similar to
those presented in this section, substituting a NSLOOKUP command for
DADMIN.
• We recommend that you separate the primary and backup DHCP servers
by some distance, perhaps placing them in separate offices. A fire, flood,
or other mishap may only affect one of the two servers if they are not both
in the same place.
• Your DHCP servers, particularly your backup server, should be secured in
a locked area. FAILSAFE.CMD contains the user name and password for
accessing your primary server in plain text. If an unauthorized person
gains access to FAILSAFE.CMD, you could expose your primary DHCP
server to attack. If you cannot secure your backup DHCP server, you
should find some other way of providing FAILSAFE.CMD with the user
name and password information it needs to access the primary server.
• You may wish to relocate the FTP transfer of DHCPCD.CFG to a REXX
script running on the primary server and push the file to the backup server
periodically instead of pulling it. That way the primary server can transmit
a new copy of the file only when it changes, cutting down on network
traffic. The backup server would then be running FTPD instead, further
reducing the load on the primary server. Just make sure you manually
copy the latest DHCPCD.CFG to the backup server before you start
running FAILSAFE.CMD for the first time.
You could also use periodic FTP pushes to verify whether the backup
server has failed or not, then generate an alert at the primary server if the
file transfer does not succeed.
• One backup DHCP server could conceivably monitor more than one
primary DHCP server with some careful additions to FAILSAFE.CMD.
• Resetting the IP address of the backup server to match the address of the
primary server can be dangerous if the primary server has not actually
failed. If your network hardware supports it, you may wish to have
Chapter 9. Reliability
379
FAILSAFE.CMD literally disable the port used by the primary server in the
event a failure is detected, to make it impossible for the primary server to
rejoin the network.
9.4 AIX features
You may wish to opt for AIX-based servers as your need for a large, reliable,
and stable network continues to grow.
IBM AIX can be made fault-tolerant by including clustering technology, such
as High Availability Cluster Multi-Processing (HACMP). The redbook titled
HACMP Enhanced Scalability Handbook, SC24-5328 is a good source of
information on this technology.
An example implementation of HACMP would include two AIX systems
running HACMP. Both have access to shared disks that contain DHCP and
DNS data files. HACMP failover scripts on the secondary machine are written
to start DHCP and DNS servers in the event of a failure on the primary
system.
9.5 Shadow IPserver features
Shadow IPserver is a product that is available as a software-only solution, or
combined with a rugged, fault tolerant hardware platform. IPserver can be
expanded into a cluster of peer servers to implement redundant DHCP,
dynamic DNS, and NBNS services. The IPserver peers are independent
servers that may be installed alongside each other or in different locations.
Each server coordinates its activities with the other across the network.
The IPserver cluster architecture allows real-time data replication. Fully
redundant data distribution to multiple IPservers provides high reliability,
tolerance of hardware failure, and the ability to balance the client load across
multiple servers. Please refer to Part 6.4, “Dynamic IP with Network
TeleSystems Shadow IPserver” on page 293 for more information on Shadow
IPserver.
380
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 10. Performance
If you're looking to boost the performance and capacity of your TCP/IP
network, this chapter should help. This chapter examines how to set lease
times and minimize broadcast traffic, among other issues.
10.1 Leases
The lease time implemented in your network will vary depending on how your
network is configured, and by the nature of the clients that are attached to it.
Here, we examine how you could determine what the lease time should be on
your network.
The lease time is the time that a DHCP client has to use the parameters
supplied to it by a DHCP server. The time itself is one of the parameters
passed. More information on the mechanics of DHCP leases can be found in
Section 3.2, “The DHCP lease” on page 114.
10.1.1 Choosing a lease time
There are no hard and fast rules for setting lease times. They can vary from
site to site according to usage patterns, goals, service levels for the DHCP
servers, and WAN stability.
In determining your optimum lease time, ask yourself these questions:
• Do I have more users than IP addresses?
For example, you may have many mobile users coming and going from
your location. Each user that leaves your site can still have a valid lease.
• How many clients do I have?
If you have a fairly short lease time and a large number of clients, you may
begin to impact the performance of your DHCP server.
• How long will it take to repair or restart a failed DHCP server?
If a DHCP server fails, clients trying to obtain a lease will encounter
problems. Clients already active will have problems when their T2 timer
expires.
• Will I need to change IP parameters at the client?
Are there parameters that are likely to change that are needed by the
clients to work correctly, such as a new default router?
• How reliable are WAN connections?
© Copyright IBM Corp. 1998, 2000
381
The affects of a WAN outage on a client are the same as if the DHCP
server itself has failed. The server may experience heavy loads when a
failed WAN connection is reestablished and all the remote clients try to
obtain leases.
• Will I need to reclaim addresses?
Will you need to get back IP addresses so they can be statically assigned
to new hosts, such as a new router.
Recommendation
Make the lease time twice as long as any potential outage of the DHCP
server.
Remember that the longer the lease, the longer it takes for changes in
DHCP options to reach the clients.
10.1.1.1 Example lease times
Here’s a list of example lease times, with the possible reasons you could use
them:
15 minutes
Allows you to maximize available addresses when you have a
dynamic environment with lots of mobile users and, therefore, a
shortage of IP addresses.
6 hours
Allows you time to repair a failed DHCP server. (three hours)
12 hours
Allows you to distribute new IP parameters and reclaim an IP
addresses overnight.
24 hours
Allows clients to continue working in the morning if a DHCP
server has failed overnight.
3 days
Appears to be Microsoft’s default value, so is used by many
locations.
6 days
Allows clients to continue working on Monday morning if a
DHCP server has failed over a weekend.
1 year
If a client has not used the address for a year, they have
probably left, and the address can be reused.
10.1.2 DHCP server differences
Based on the lease time that you specify, the DHCP server automatically
calculates a renewal interval and a rebinding interval as suggested by the
DHCP RFC. The renewal and rebinding options are described more fully in
3.2.1, “Getting the lease” on page 114.
382
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Some server implementations allow greater control over these intervals by
letting you specify them individually. If your server supports this function, you
gain a greater latitude in updating client configuration. For example, you can
specify a lengthy total lease time to protect your clients from DHCP server
outages, yet specify a lease renewal time that is much shorter, so that
updated configuration options are distributed in a more timely fashion.
In our LAB environment we found that AIX, OS/2 and ShadowIP DHCP
servers allowed us the flexibility to specify the three intervals independently.
The Microsoft NT DHCP server only allows entry of the total lease time. It
calculates the values for the renewing and rebinding options at 50% and
87.5% of the total lease time, respectively.
10.1.3 Multiple leases
With two or more DHCP servers on a network, clients that are moved around
(for example, laptops) can end up with multiple and redundant leases.
Consider a local network with two DHCP servers and a remote site also
served by those servers. A mobile client initially connects to the local network
and receives a lease from one of the two servers. It is then moved to the
remote network without releasing the lease. Because this new location is on a
different subnet, the client’s current IP address is not appropriate. It will then
receive another address from the server, while still holding a valid lease on
the first address.
If the client is then moved back to the original network, it will attempt to use
the IP address from the second lease, which is not valid on this subnet. At
this point the server might offer the client its original address, or the client
could get a lease on a third address.
This problem can be eased by using only one DHCP server per network or
location and also by using short lease times. If in your network, DHCP
servers are maintained by individual departments, offices, or individuals with
their own small address pools, they can find that their addresses are being
used by any device on the LAN that has been configured to use DHCP.
Another potential problem with using a decentralized method of DHCP
administration is the way some clients behave when switched between
networks. In Chapter 3 we discussed a scenario where a Windows client took
over 15 minutes to get a lease after being moved between networks. Please
refer to Part 3.4.4, “DHCPNAK” on page 126 for details of this common
problem.
Chapter 10. Performance
383
In our test network we implemented a DHCP solution that was able to
differentiate our desktop clients from our laptop clients and configured the
laptops with much shorter lease times. Please refer to Part 6.4.6, “DHCP
pools” on page 307 for information on our experience with the Shadow
IPserver.
10.2 Monitoring and troubleshooting
Most TCP/IP implementations come with a few standard tools that can be
used in monitoring and debugging your IP network. Alternatively, your could
use third-party applications and dedicated analyzers to do this more
thoroughly.
The following commands are included with most implementations of TCP/IP
and can be used to check connectivity. The commands have been
documented using TCP/IP Version 4.21 for OS/2. Their usage may vary with
other operating systems.
10.2.1 The PING command
By sending ICMP echo requests to specific IP addresses, PING (short for
Packet Internet Groper) is used to verify connections between hosts.
PING Syntax
PING [-?drv] host [size [packets]]
Where:
-?
Displays the syntax of the PING command
-d
Turn debug on
-r
Bypass the normal routing tables
-v
Verbose output. Include all ICMP packets received
host
A host’s name, IP address or broadcast address
size
The size of data portion of the packet
packets
The number of Echo Request packets to send
For example:
[C:\]PING banquet 56 2
PING banquet.AUSTIN.COOKING.NET: 56 data bytes
64 bytes from 192.168.6.1: icmp_seq=0. time=0. ms
384
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
64 bytes from 192.168.6.1: icmp_seq=1. time=0. ms
----banquet.AUSTIN.COOKING.NET PING Statistics---2 packets transmitted, 2 packets received, 0% packet loss
round-trip (ms) min/avg/max = 0/0/0
10.2.2 The TRACERTE command
The TRACERTE command is used to trace the route an IP packet takes to reach
a host on a remote subnet. It can tell you where and why a route is lost by
displaying information received from the intervening routers between you and
your destination. This command is available on AIX systems as traceroute,
and on Windows NT systems as TRACERT.
TRACERTE works by sending a series of PINGs to the target host. The first PING
it sends is coded with time to live (TTL) value of zero. The term TTL actually
refers to a number of router hops, not time. When this datagram encounters a
router along the path to the target, the router responds with an ICMP
message that says the TTL expired in transit. In this manner, the first router
between you and the target host is identified. The TRACERTE program
continues to send PINGs, incrementing the TTL value by one each time, so
that every router in the path is identified. The error indications displayed by
TRACERTE are derived from the ICMP messages it receives from routers in
the network.
TRACERTE Syntax
TRACERTE [-dnrv] [-w wait] [-m max_ttl] [-p port#] [-q nqueries]
[-t tos] [-s src_addr] host [data_size]
Where:
-d
Turn debug on
-n
Display IP addresses instead of host names
-r
Disables routing of IP packets
-v
Turns verbose on
-w wait
Wait time in wait seconds between packets
-m max_ttl
Maximum time-to-live in max_ttl hops
-p port#
Destination port number port#
-q nqueries
nqueries number of probes for each TTL
-t tos
tos type of service
Chapter 10. Performance
385
-s src_addr
Source IP address src_addr
host
A host’s name or IP address
data_size
The number of bytes of data used
For example:
[C:\]tracerte 192.168.7.10
traceroute to 192.168.7.10 (192.168.7.10), 30 hops max, 38 byte packets
1 192.168.6.1
(192.168.6.1) 0 ms 0 ms 10 ms
2 brie
(192.168.7.10) 0 ms 0 ms 0 ms
TRACERTE interprets ICMP messages received from routes and displays them
in using these indicators:
!
Port unreachable, connection refused
!N
Network unreachable, no route to host
!H
Host unreachable, no route to host
!P
Protocol unreachable, connection refused
!F
Message too big
!S
No route to host
For example:
[C:\]tracerte 192.168.10.1
traceroute to 192.168.10.1 (192.168.10.1), 30 hops max, 38 byte packets
1 192.168.6.1
(192.168.6.1) 10 ms 10 ms 0 ms
2 192.168.6.1
(192.168.6.1) 0 ms !H 0 ms !H 0 ms !H
This indicates that our default router, 192.168.6.1, has no routing information
to reach the host 192.168.10.1. While these indicators are somewhat cryptic,
understanding their meaning can be quite valuable to your problem
determination efforts. They give you a perspective of the network from a
viewpoint within a router that may be many hops away.
10.2.3 The IPTRACE command
The IPTRACE command is used to trace all packets received by and sent from a
network interface. All data collected will be written to the IPTRACE.DMP file
in the current directory. Depending on the platform, the dump file may be
stored in a non-editable format. In OS/2, the IPFORMAT command is used to
view the dump file. The text output from the IPFORMAT can be redirected to a
file. Additionally, entering the command IPFORMAT -X will produce a binary file
in Network General Sniffer format.
386
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
In AIX, the IPREPORT command is used to examine the trace data.
IPTRACE syntax
IPTRACE [-i] [interface]
Where:
-i
Specifies that only IP packets should be traced
interface
Specifies an interface to be traced
10.2.4 The ARP command
ARP is a mechanism to dynamically map IP addresses to the MAC addresses
of other network adapters in the same local subnet and then cache them in
memory for future reference. You can use the arp command to display and
manually maintain the ARP cache.
arp syntax
arp [-afds?] hostname [hardware_addr] [temp|pub]
hostname A host’s name or IP address
-?
Show the online help
-a
Print all arp table entries
-f
Flush all arp table entries
-d
Delete arp table entry for the host hostname
-s
Add arp table entry for the host hostname
temp
Time out this entry if it is not used
pub
Reply for other host
10.2.5 The NETSTAT command
The NETSTAT command displays the network status of the local workstation. It
supplies information about routing, TCP connections, UDP statistics, IP
statistics, memory buffers, and sockets.
NETSTAT syntax
NETSTAT [ -? ] [ -acghilmnprstu]
Chapter 10. Performance
387
Where:
a
Displays addresses of network interfaces
c
Displays ICMP statistics
g
Displays IGMP statistics
h
Displays the resolved host name
i
Displays IP statistics
l
Displays information about the socket that is listening
m
Displays information about memory buffer usage
n
Displays information about LAN interfaces
p
Displays the contents of the ARP table
r
Displays the routing tables
s
Displays information about sockets
t
Displays information about TCP connections
u
Displays UDP statistics
For example, to dump the current routing table:
[C:\]NETSTAT -r
destination
router
netmask
default
127.0.0.1
192.168.6
192.168.7.10
192.168.9.10
0.0.0.0
255.255.255.255
255.255.255.0
255.255.255.255
255.255.255.255
192.168.6.1
127.0.0.1
192.168.6.10
192.168.6.1
192.168.6.4
metric flags intrf
0
UGP
0
UH
0
UC
0 UGHW3
0 UGHDM
lan0
lo
lan0
lan0
lan0
10.2.6 The HOST command
The HOST command is used to ask a DNS server to resolve hostnames to IP
addresses and vice versa.
HOST syntax
HOST <hostname>
hostname A host’s name or IP address.
The HOST command is a quick and easy tool to use on OS/2 and AIX systems,
but is not available in the Windows NT environment. Here you can use the
NSLOOKUP command.
388
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
10.2.7 The NSLOOKUP command
NSLOOKUP is used to query DNS servers in either an interactive or
non-interactive fashion.
NSLOOKUP syntax
NSLOOKUP [-options]
(Interactive mode using default server)
[-options] -server
(Interactive mode using server)
[-options] host
(Look up host using default server)
[-options] host server
(Look up host using server)
Where [options] are:
all
Print options, current server and host.
[no]debug
Print debugging information.
[no]d2
Print exhaustive debugging information.
[no]defname
Append domain name to each query.
[no]recurse
Ask for recursive answer to query.
[no]search
Use the search list.
[no]vc
Always use a virtual circuit.
domain=name
Set default domain name to name.
port=x
Use TCP/IP port number x.
srchlist=n1[/n2 /.../n6]
Set domain to n1 and search list to n1, n2, etc.
root=name
Set root server to name.
retry=x
Set number of retries to x.
timeout=x
Set initial time-out interval to x seconds.
querytype=x or type=x
Set query type, for example: A, ANY, CNAME, NS, PTR..
class=x
Set query class to one of IN (Internet), CHAOS,
HESIOD or ANY.
10.2.8 Other utilities
Here (briefly) are some tools that you could consider using to monitor and
troubleshoot your network.
Chapter 10. Performance
389
10.2.8.1 Netfinity
Netfinity Manager and Client Services for Netfinity Manager is not a tool for
administering TCP/IP. Rather, it is a systems management tool that allows
you to monitor, manage and configure both local and remote systems.
Netfinity can be used over a variety of transport protocols such as TCP/IP,
SNA and NetBIOS.
The Systems Monitor Service, part of Netfinity, allows you to collect statistics
from active systems. For example CPU, memory, disk and print usage can be
monitored. You can also use the Systems Monitor Service to collect TCP/IP
statistics.
See Appendix B.3, “Netfinity” on page 447 for information regarding running
Netfinity in a DHCP environment.
10.2.8.2 Third-party alternatives
There are many third-party network monitoring and analyzing tools available.
These can range from software that can be run on existing PCs on the
network, to dedicated devices that attach directly to the network.
Current examples include:
• Fluke Corporation supplies several diagnostic solutions such as
OneTouch, Enterprise LANMeter and Network Inspector.
• Hewlett Packard has its NetMetrix range.
• Network Associates has available the software-based Sniffer Basic and
the dedicated Sniffer Pro LAN.
10.3 Troubleshooting TCP/IP networks
Sooner or later, you will probably encounter problems on your network. How
quickly you can resolve these problems depends on your approach to
troubleshooting. Here, we examine ways to debug your network using the
commands that come standard with most implementations of TCP/IP.
10.3.1 Prerequisites for troubleshooting
If you encounter connectivity problems with your network, it may take many
steps to find the problem, and hence the solution. When debugging a network
problem, keep the following prerequisites in mind.
390
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
10.3.1.1 Understand TCP/IP
To solve an IP related problem, it goes without saying that you have to have a
high understanding of TCP/IP. TCP/IP is an open, multi-vendor protocol. As
such, there is a lot of publicly available information – see your local
bookstore.
10.3.1.2 Know your environment
As a network administrator, you should know what equipment and systems
have been placed onto the network. Not only should you know what
equipment is out there, you should know and understand their role within the
greater whole.
10.3.1.3 Any information can help
Any symptom can be a clue to currently occurring problem, even if at first
glance it appears to be unrelated.
10.3.1.4 Don’t believe what people say
Don’t believe what people have told you until it has been verified. If the
network is able to provide telnet connectivity, it is a good way to get reliable
information from a remote portion of your network.
10.3.1.5 You can be limited by the person on-site
Your current skill level can be attenuated by that of the person on the other
end of the phone. Unless you can talk them though your diagnostic
procedures, diagnosis may not start until you can actually get on-site.
10.3.1.6 Don’t hesitate to ask for help
Today’s network environment is increasingly complex and constantly
changing. Network administration and support staff are faced with the need to
constantly refresh their skills while being responsible for day-to-day issues as
well. If you encounter a network issue that is consuming an excessive amount
of your resources, you may benefit from calling in some expertise from
outside of your organization. IBM Global Services for example, can offer
assistance from skilled and experienced professionals in a wide variety of
networking areas. The URL for the IBM Global Services Web site is:
http://www.ibm.com/services/
10.3.2 A bottom-up approach
Because protocols are divided into several layers, and each layer’s
connectivity depends on the layer beneath it, it is therefore reasonable to
start diagnosing network problems from the bottom and work your way up.
Chapter 10. Performance
391
OSI
7
Application
6
Presentation
5
Session
4
TCP/IP
Application
4
Transport
Transport
3
3
Network
Internet
2
2
Data Link
1
1
Network
Interface
Physical
Figure 267. OSI and TCP/IP comparison
Figure 267 compares the seven-layer OSI model to a layered view of TCP/IP.
Although the differences in the stacks are obvious, a layered approach to
understanding and troubleshooting is beneficial in both. Each layer is
explained from the bottom to the top.
10.3.2.1 Layer 1 physical layer
The physical layer represents the media used within the network, such as
fiber optic, twisted pairs of copper, or coaxial cable. It is only responsible for
the transmission of data across the physical network.
10.3.2.2 Layer 2 data link layer
The data link layer represents a communication link between two systems.
Network types such as Ethernet, token-ring and FDDI are covered by this
layer. The data processed in this layer is often called a frame.
In debugging the data link layer, you will probably use the arp command as
documented on page 387. Keep the following in mind:
• When arp fails, you cannot communicate.
• When arp succeeds, you don’t have any hardware problems.
392
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• When arp succeeds, but you still can’t communicate, then you have a
problem in a higher layer.
The arp command cannot be used to diagnose higher layers, but applications
that use the higher layers (such as PING) can assist in diagnosing lower layers.
In using arp for testing, use the following procedure:
1. Clear the ARP cache with the an arp -f command.
2. Check that the cache is empty with an arp -a command.
3. Try to PING the host you are trying to connect to.
4. Then check the ARP cache again. If the PING was successful, and the host
you PINGed is on the same subnet, there will be an entry for the host in the
table. If the host is on another subnet, there should be an entry for the
router that would be used. (Note in the following example that an ARP
entry for our DNS server with the address 192.168.6.5 was also added.)
[C:\]arp -a
ARP table contents:
interface
hardware address
lan0
lan0
0 :6 :29:b3:e :ed
8 :0 :5a:ce:ea:cb
IP address
minutes since
last use
192.168.6.1
0
192.168.6.5
0
If the PING was unsuccessful, and the host is on the same subnet, try to
PING other devices on the subnet. If the host being PINGed is on a different
subnet, try PINGing the default router. By doing this, you can receive some
indication as to where to next proceed.
10.3.2.3 Layer 3 network layer
The network layer represents communication between multiple systems. In
TCP/IP, IP and ICMP are network layer protocols, but IP is the only protocol to
carry user data. The data passed through this layer is often called a packet.
(An IP packet is sometimes called an IP datagram). The IP address is used to
identify various systems within the network.
This layer also allows for the connection between two systems by passing
through other systems, through a process known as routing.
In debugging the network layer, there are several commands that can be
useful.
Use the PING command
The PING command is documented on page 384. When using PING, keep the
following in mind:
Chapter 10. Performance
393
• When PING fails, you cannot communicate by using the IP protocol
• When PING succeeds you don’t have hardware or network configuration
problems.
• When PING succeeds, but communication still fails, there is an application
problem.
If you are cannot connect to a host, try the following PING sequence to help
determine where the problem lies:
1. PING the IP address 127.0.0.1 by issuing the command:
ping 127.0.0.1
As discussed in 1.2, “IP addresses” on page 2, 127.0.0.1 is the loopback
IP address. If you can successfully PING this address, then you have
proven that the IP stack on your host is working correctly.
2. PING your own IP address. A PING to your own address will physically
transmit out onto the network. If the PING is successful, then you know that
your network adapter is working.
3. PING the IP address of your default router. This will determine if your
default router is connected to the network, but not that it is working
correctly. It can also show if your subnet mask is correct.
4. PING an IP address of a host on the same subnet of the host you want to
connect to. This will determine if all routers between you and the host you
want to connect to are working correctly.
5. PING the IP address of the host you want to connect to.
In the above procedure, we have used IP addresses only because any
problems in communicating with your name server, or mis-configurations in
your name server, can cause connectivity problems in general when you are
trying to connect by name.
In the following example, we are trying to PING the machine called hotdog:
[C:\]PING hotdog 56 2
PING hotdog.AUSTIN.COOKING.NET: 56 data bytes
----hotdog.AUSTIN.COOKING.NET PING Statistics---2 packets transmitted, 0 packets received, 100% packet loss
As you can see, the PING command itself has failed, but the host name was
successfully resolved by the name server. This proves that the name server is
working, although the address supplied by the name server may not
necessarily be correct.
394
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The PING command is also an excellent means of identifying frame size
issues in your network. Most implementations of the command allow you to
specify the size of the echo packet.
Use the TRACERTE command
If, in step 4, if you could not PING any hosts on the remote subnet, you can use
the TRACERTE command as documented on page 385 to determine where and
why data is being lost.
Use the NETSTAT command
Use the NETSTAT command as documented on page 387. In OS/2, use
NETSTAT -i to check IP statistics, and NETSTAT -c to check ICMP statistics.
Figure 268 on page 395 shows partial output from the NETSTAT -i command.
[C:\]NETSTAT -i
total packets received 5548
checksum bad 0
packet too short 0
not enough data 0
ip header length < data size 0
ip length < ip header length 0
fragments received 0
frags dropped (dups, out of space) 0
fragments timed out 0
packets forwarded 0
packets rcvd for unreachable dest 0
packets forwarded on same net 0
Unknown/Unsupported protocol 2
requests for transmission 4974
lost packets due to no bufs, etc 0
output packets discarded because no route could be found 0
input packets delivered successfully to user-protocols 5546
input packets with an unknown protocol 0
output packets successfully fragmented 0
output fragments created 0
fragmentation failed 0
successfully assembled packets 0
Packets received with version !=4 0
Raw ip packets generated 193
Figure 268. Output from NETSTAT -i
Look for unusual counters. If output packets discarded because no route could
be found is not zero, you have a routing problem somewhere. If NETSTAT -c
gives a high Source Quench count, the host you are talking to (or the routers
you are talking through) may be overloaded.
In Windows NT, use the netstat -s command to see statistics for all protocols,
and for AIX, use netstat -p <protocol> to see statistics for a specific protocol,
such as TCP or UDP.
Chapter 10. Performance
395
10.3.2.4 Layer 4 transport layer
This layer represents a connection between two processes. Any system can
have multiple processes running on it, and TCP and UDP are the protocols
used to achieve this connection. TCP data passed through this layer is often
called a segment; UDP data is known as a datagram.
This layer also provides functionality for flow control and reliability (including
retransmission). TCP provides these functions, while UDP does not.
Use the NETSTAT command
Use the NETSTAT command as documented in 10.2.5 on page 387. Use
NETSTAT -t to check TCP statistics and NETSTAT -u to display UDP statistics.
Additionally, use NETSTAT -s to check socket information. Figure 269 on page
396 shows part of the output from the NETSTAT -t command.
[C:\]NETSTAT -t
TCP STATISTICS
connections initiated
connections accepted
connections established
embryonic connections dropped
conn. closed (includes drops)
segs where we tried to get rtt
times we succeeded
delayed acks sent
conn. dropped in rxmt timeout
retransmit timeouts
persist timeouts
keepalive timeouts
keepalive probes sent
connections dropped in keepalive
total packets sent
data packets sent
data bytes sent
data packets retransmitted
data bytes retransmitted
ack-only packets sent
window probes sent
packets sent with URG only
window update-only packets sent
2
1
2
0
4
10
8
8
0
0
0
0
0
0
23
6
92
0
0
12
0
0
1
Figure 269. Output from NETSTAT -t
Again, look for unusual counters. If packets received with ccksum errs is not
zero, you are experiencing errors somewhere on your network.
10.3.2.5 Layer 5 session layer
The session layer provides dialog sessions such as full- and half-duplex and
synchronization points in the dialog. Within TCP/IP, there is no precise
session layer, although some of its functionality is provided within TCP.
396
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
10.3.2.6 Layer 6 presentation layer
The presentation layer provides for common data presentation between
applications. Within TCP/IP, there is no precise presentation layer.
10.3.2.7 Layer 7 application layer
The application layer represents the application entity, usually an application
program.
When diagnosing DNS issues, there are two commands that are useful.
Use the HOST command
The HOST command, as documented on page 388, is used to query a DNS
server and can be used to verify that your DNS configuration is correct. HOST
can only be used for name to address and address to name resolution.
Use the NSLOOKUP command
The NSLOOKUP command, as documented on page 389, is used to query a DNS
server and can be used to verify that your DNS configuration is correct.
You should also read the TCP/IP Command Reference in the TCP/IP
Information folder of Warp Server.
This section provides an overview of how to use NSLOOKUP. The information
presented here should work for both OS/2 and many UNIX versions of
NSLOOKUP. All Windows-based versions of nslookup that we have seen are GUI
programs and have different (and varying) usage.
NSLOOKUP is a resolver that sends queries to name servers. This is the same
task that your browser has to do when locating a particular Web page (if you
entered the server name rather than the server IP address). By default,
NSLOOKUP performs recursive queries. See Figure 270 on page 398 for an
example of how a name server resolves a recursive query. Resolvers make
recursive queries, which cause the name server the most amount of work.
Name servers themselves make non-recursive queries. You can configure
NSLOOKUP to do non-recursive queries as well.
Chapter 10. Performance
397
Figure 270. Name resolution example
NSLOOKUP is a useful while you are configuring your name server(s) and also
for troubleshooting certain issues after configuration. For example, pretend
you have configured one of your servers to provide zone transfers to only a
specified group of machines (for example, by using the secure_zones option). It
is probably prudent to use nslookup to try a zone transfer (using the ls
sub-command) from a machine other than the trusted machines, rather than
to wait for crackers to test your configuration for you. Most name servers, you
will find, will not pass you zone information for security reasons.
NSLOOKUP can be run interactively, or, if you have a single query, you can type
the query from the command line. When you start NSLOOKUP interactively, it is
useful to enter the command set all to remind you of your default settings
(particularly your default name server, domain, and search list, if any). To
finish an interactive session, type exit or press CTRL+Z (end of file on OS/2)
or CTRL+D (end of file on UNIX). For extra help during an interactive session,
type ? or help at the > prompt.
NSLOOKUP can be configured for recursive and non-recursive queries by using
the set recur and set norecur subcommands. You can start the program so
that it uses the default name server, or you can direct it to use a specific
name server for your session. The file %ETC%\RESOLV2 will contain the
name of your default name server (if any), and your default domain. An
398
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
example of %ETC%\RESOLV2 is shown below for a machine on our
armonk.cooking.net subnet.
domain armonk.cooking.net
nameserver 192.168.6.10
You could include the line options debug, which turns on debugging for
NSLOOKUP. When you first start using NSLOOKUP, it may be useful to have debug
on to better understand what is happening. With debug on (or d2), you will see
about as much DNS information as you would see if you were using a
protocol analyzer. We are not using debug here because of the amount of
information it creates. We suggest that you use debug, though, while you are
becoming familiar with NSLOOKUP.
The NSLOOKUP syntax on page 389 shows how to specify a particular name
server to use. For example:
NSLOOKUP - 192.168.7.10
Server: fajita.austin.cooking.net
Address: 192.168.7.10
> exit
This starts NSLOOKUP using the FAJITA name server, rather than our default
name server merlot.armonk.cooking.net. This is useful when your default
server is not available due to network or other problems. You will know there
is a problem with your default name server if NSLOOKUP exits immediately with
an error message indicating:
• Timed out
• No response from server
• Connection refused
• Server failure
• Default servers not available
If you do get errors like these, you probably want to start NSLOOKUP using a
different name server until you can determine what is wrong that the default
name server.
Do not confuse the previous errors with errors of the type:
• Nonexistent domain (host or domain is not known to name server)
• Query refused (often because of security settings at name server)
• Format error (possible error in NSLOOKUP or network packet error; the name
server thinks request packet is an improper format)
Chapter 10. Performance
399
• No records (of the type you are requesting; change your querytype to any
and try again)
You can also redirect some NSLOOKUP output to a file. For example, when in
interactive mode if you wanted to create a file of all mail exchanger (MX)
records (that is, mail hosts) for the domain armonk.cooking.net, you would do
the following commands (the t is for recordtype):
nslookup
Server: fajita.armonk.cooking.net
Address: 192.168.6.10
> ls -t MX armonk.cooking.net > mailhost.txt
> view mailhost.txt
While records are still being read, you will see a series of # signs going
across the screen; so you know something is happening. The view command
shows you the file as if you were using the more command (which is what view
uses). If you want to halt a long command, you can type CTRL+C during the
command, which will bring you back to the > prompt. For example, if you
entered the ls -t MX command on a large Internet domain, it could potentially
generate pages of information.
> ls -t any armonk.cooking.net
This would show all records for the domain, which would, of course, give even
more information than just the MX records. The record types that you can show
individually are:
400
A
Host name to IP address records.
ANY
All records available from name server.
CNAME
Canonical names for aliases. For example, we use ns-updates as
an alias for the canonical name of fajita.armonk.cooking.net.
Usually, you will see mail or www as an alias for the fully qualified
domain name of the server the mail server or Web server runs on.
HINFO
Host information.
KEY
Public key information.
MB
Mailbox information.
MINFO
Mailbox or mail list information.
MX
Mail exchanger.
NS
Name server.
PTR
IP address to host name records.
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
SIG
List records with expiration time and signature information such as
A and KEY records.
SOA
Domain’s start of authority.
TXT
Text records.
UINFO
User information. For example, it will sometimes contain contact
information of responsible person for a particular server.
WKS
Well-known services that the host advertises.
Note
The record types are discussed in more detail in 1.6.5, “Resource record
types” on page 23.
When you enter the querytype to nslookup sub-commands, remember that
they are case insensitive. The ls sub-command causes zone transfers, and
some hosts will not allow that (for security or because of the extra workload
on the name server and network).
The set command is used when in the interactive mode to change default
settings. For example, it can be useful to do a set debug or set d2 (both of
which turn on debugging) so that you can see how the name resolution is
occurring when you make a query. You can change the default querytype for a
single command. For example, to show all name server records for
armonk.cooking.net domain, enter:
[C:\]NSLOOKUP -query=NS armonk.cooking.net
Server: fajita.armonk.cooking.net
Address: 192.168.6.10
armonk.cooking.net nameserver = fajita.armonk.cooking.net
armonk.cooking.net nameserver = ns-update
fajita.armonk.cooking.net internet address = 192.168.6.10
ns-update internet address = 192.168.6.10
Your default record query type can be set to NS as opposed to the default of
any records:
[C:\]NSLOOKUP -query=NS
Default Server: fajita.armonk.cooking.net
Address: 192.168.6.10
>
Chapter 10. Performance
401
We are now in the interactive mode and can give more commands. To change
our default name server:
> set root=192.168.7.10
>
To change our default querytype back to any:
> set q=any
We can abbreviate most sub-command arguments using the first unique
character pattern of the name. Now we would be using the name server for
our austin.cooking.net subnetwork. This could become confusing because
our default domain is still armonk.cooking.net, so we might want to use a
search list to include both domains:
> set srchlist=austin.cooking.net/armonk.cooking.net
>
Now, when the name resolution occurs, first austin.cooking.net will be
appended to the name we are looking up, then armonk.cooking.net will be
appended, then the name will be searched by itself with no appending. For
example:
> jalapeno
Name resolution occurs as follows. When defname and search are both on
(the default), NSLOOKUP uses the search list and successively appends each
name in the list to any host query not ending in a dot until the name resolution
is successful (jalapeno.austin.cooking.net and then
jalapeno.armonk.cooking.net would be searched). If no attempt is successful,
NSLOOKUP tries the user-entered query as-is. If you do not want the search list
to be used you can append a dot (.) to your command as follows:
> jalapeno.austin.cooking.net.
and then the name is used as is with no appending. Note that the name
resolution stops when an address is found. If search is off and defname is on,
NSLOOKUP does not use the search list, but appends only the domain name
(which is the first entry in the search list). Also, NSLOOKUP will only append this
domain name to user queries that do not have dots in the name, (rather than
names which do not end in a dot, which was the case when search was on).
If defname is off, no domain is appended to the query. In this case, setting
search/nosearch has no effect.
402
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
10.4 Tuning TCP/IP networks
Tuning is an ongoing function in networking. Bandwidth and performance are
constantly monitored using fully implemented state of the art network
management methods. New applications and hosts are introduced to the
environment only after their network impact is fully understood. Change is
carefully controlled and is normally transparent to the end user.
10.4.1 An approach to tuning your network
If you think the network environment described above sounds utopian, you’re
right. In practice, we most often see network tuning efforts initiated as a result
of a network performance issue. Therefore, that is the approach we’ll focus on
in this section.
10.4.1.1 Gather information
Gather as much information as possible so that potential causes of your
problem (in this case poor performance) can be found. Some questions to ask
yourself are:
• Do all applications suffer from poor performance?
• Do all operations within an application suffer from poor performance?
• Do all systems within the network suffer from poor performance?
• When did the problem begin?
• Is the problem intermittent?
• Are there any error messages on either the console or logged to a file
somewhere?
10.4.1.2 Understand your environment
You have to know and understand the environment that you are dealing with.
This includes all products – both hardware and software – and protocols that
actively use the network.
Check that all products are being used within specification. For example,
most networking environments specify the maximum number of workstations
and maximum cable lengths that can be used in a network segment. If
additional workstations have been added to the network, this maximum may
have been exceeded.
10.4.1.3 Is it really a problem?
Are your users expecting too much? If they are used to loading files from a
local file server, they may start to complain that it’s too slow when
Chapter 10. Performance
403
downloading from the Internet. A T1 connection is not going to give the same
throughput as a 16 Mb token-ring.
10.4.1.4 Is it really a network function?
Don’t forget that memory usage, disk, CPU, and I/O activity on both the
source and target hosts can influence performance. Before you delve deeply
into the network, confirm that these resources are not constrained on your
hosts.
10.4.1.5 What do you want to achieve?
Are you trying to achieve a better response time, or greater throughput in
general? Tuning the response time may entail tuning a particular application
on a host, whereas to get better throughput, you might have to visit every
workstation, host and router.
10.4.2 TCP/IP tuning parameters
When tuning your network, there a several parameters that you may be able
to manipulate to enhance network performance. The ability to change these
parameters is dependent on the implementation of TCP/IP that you are using.
10.4.2.1 MTU and fragmentation
The Maximum Transfer Unit is a very important parameter. As TCP/IP works
in almost all situations without modifying the default MTU, the MTU is not
often tuned. By changing the MTU size, you can minimize the fragmentation
that occurs on a packet as it travels through your IP network.
Consider the following:
• When you are sending small packets, and there is no fragmentation:
• More packets have to be sent; so, there is an increased number of I/O
operations to the network. This adds overhead.
• There is lower throughput.
• When you are sending large packets, and there is no fragmentation:
• Fewer packet have to be sent; so there is a reduced number of I/O
operations to the network. This reduces overhead.
• There is higher throughput.
• When you are sending large packets, and there is fragmentation:
• More packets have to be sent because the original data has been
fragmented. This increases I/O operations to and from the network.
• There is lower throughput.
404
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• More packets must be transmitted to recover from lost data packets.
Given the above, we can see that we should set the MTU size to be as large
as possible so that a packet travelling through the network does not become
fragmented.
You can use the PING (see page 384) and TRACERTE (see page 385) commands
to determine the bottlenecks that occur on your network.
Use TRACERTE to determine the IP addresses of the routers between you and
the host you are connecting to. Then PING each IP address found with a range
of packet sizes (use the same range for each address) and graph the results.
Any bottlenecks with larger MTU sizes should be readily apparent.
Remember to make the MTU size on your workstation larger than the packet
sizes you are using so that fragmentation does not occur before the data
reaches the network.
To change the MTU size, you would usually use the IFCONFIG command,
although some implementations of TCP/IP may use configuration files to hold
the MTU definition.
For example, under OS/2 you could issue the commands:
IFCONFIG lan0 down
IFCONFIG lan0 mtu 4400
IFCONFIG lan0 up
This will set the MTU size to 4400 bytes. Now issue the command:
NETSTAT -n
This will confirm that the change has been made.
Installations that use DHCP can easily change the MTU size for multiple
hosts. It is BOOTP\DHCP option number 26.
10.4.2.2 TCP maximum segment size
As stated previously, a unit of TCP transmission is known as a segment. IP
and UDP datagrams each have a maximum length of 65,536 bytes. A TCP
segment also has a maximum length, but it is determined when two systems
connect. The maximum segment size or MSS is negotiated by both systems
to guarantee connectivity.
The MSS is calculated using the MTU as follows:
1. Calculate variable A.
Chapter 10. Performance
405
A = MTU - (TCP header size + IP header size)
2. Calculate variable B.
B = Socket Receive Buffer size / 2
3. The smaller of A and B is used as the effective MSS.
Now consider the following scenario. Your workstation is connected to a
token-ring network and communicates with hosts also attached to the ring. To
enhance performance, you are using a large MTU of 17960 bytes. Later, you
have to connect over a router to hosts attached to an Ethernet network. To
avoid fragmentation at the router you should set your MTU to 1500 bytes. But
now you have an inefficiency when talking to hosts connected to the
token-ring.
How do you resolve this problem? Some stacks have MTU discovery enabled,
thereby bypassing the issue. If they don’t, you can use the -mtu option with
the ROUTE command. The -mtu option allows you to configure the MTU for each
route, not just for each interface. It allows you to give a specific value to only
the traffic that is passing through a router.
For example, we have workstations on our 192.168.6 token-ring network.
They are using an MTU of 4400 bytes. If they need to connect to the
192.168.9 Ethernet network, we could issue the command:
route add -net 192.168.9 192.168.6.4 -mtu 1500
Here, we are assuming that 192.168.6.4 is the router to the 192.168.9
network. By adding this specific MTU size for traffic to the Ethernet network,
we can now communicate efficiently with both local and remote hosts. Our
MTU and MSS would are large, however, without fragmentation.
10.4.2.3 The IP queue
IP keeps a queue for incoming IP datagrams. The queue is only used for
receiving datagrams, not sending.
Since routing is done at the IP layer, the IP queue is important in a system
configured as a router because it is used when passing datagrams from one
network to another. If a router receives a burst of IP datagrams, it is possible
for the queue to overflow.
Many implementations of TCP/IP do not allow you to set the queue size. AIX,
for example, does.
To set and check the IP queue:
406
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
1. To check the current queue size, issue the command:
no -o ipqmaxlen
2. To set the IP queue length, issue the command:
no -o ipqmaxlen=x
Where x is the new size of the queue.
3. To check for IP queue overflows, issue the command:
netstat -p ip
The following is an extract from the output, with the counter we are
interested in highlighted:
...
0 path MTU discovery memory allocation failures
0 ipintrq overflows
0 with illegal source
Note
A larger queue may require more CPU time to process; so don’t make it too
large. If you must increase it, make the increments moderate.
10.4.2.4 Buffers
The buffers are temporary data storage used to store data sent between an
application and the network.
Some applications can set their buffer size (by using the setsockopt()
function), but generally you cannot modify the buffer size.
You can the buffers with the NETSTAT command.
NETSTAT -m displays details for memory buffer usage.
NETSTAT -c displays ICMP statistics. When looking at the ICMP statistics, look
for high Source Quench counters.
If an ICMP Source Quench message is received from an router, it means that
the router does not have the buffer space needed to queue the datagrams for
output to the next network.
10.5 Bandwidth efficiency
While the easiest solution to network performance problems may be to
provide a faster connection, it’s not always the most practical. After all, it
Chapter 10. Performance
407
might not be cost effective to link every remote user or office with a T3 line.
Sometimes it pays to closely examine the amount and nature of the network
traffic flowing through your network connections to see if it’s possible to use
limited bandwidth more efficiently. Frankly, some applications just don’t do
well in a WAN environment. For example, if a client and server must serially
interchange a large number of packets to complete a transaction, the total
time required from the user’s perspective can increase dramatically if there is
any WAN connection between two hosts.
10.5.1 Broadcast traffic
Be aware that your network may have multiple protocols running over it. Not
all protocols are efficient in how they use the network. Some, like NetBEUI for
example, introduce a large quantity of broadcast traffic.
Your environment may use combined bridge/routers (brouters) to interconnect
remote locations. If these connections are maintained over relatively slow
WAN links, your WAN link may be swamped with broadcast traffic. In extreme
cases, connections that have timer-critical response times (for example SNA),
may fail.
As we have seen in Chapter 5, “Integrating file and print services” on page
227, it is possible to remove NetBEUI from your WAN connections entirely by
implementing TCPBEUI. You can choose to still use NetBEUI within the local
network where it is most efficient, but your WAN connections will have been
freed up from carrying broadcast traffic.
10.5.2 RSVP
Unfortunately, TCP/IP has not had any standard method of prioritizing traffic
until quite recently. Yet, in the real world, bank transactions can be much
more important than casual e-mail. Unfortunately, most TCP/IP networks
can’t tell the difference between high priority datagrams and lower priority
ones; so the bandwidth gets split more or less evenly between competing
users.
RSVP (Resource ReSerVation Protocol) has been introduced as a way to
solve the problem. If your network application has been programmed to take
advantage of RSVP, and if the routers enroute support RSVP, then TCP/IP
datagrams can be prioritized according to their relative importance.
Datagrams for bank transactions can be tagged by the application as high
priority, while e-mail applications may tag their datagrams as lower priority.
Intermediate network devices, such as routers, can then handle the traffic
408
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
more intelligently, giving more bandwidth to the higher priority traffic as
needed.
There’s one major downside to RSVP, at least at this point in history: Many
applications and network devices do not yet support this protocol extension.
Hopefully the widespread adoption over time of IPv6 will help speed the
implementation of RSVP as well. Nonetheless, if you have control over your
own TCP/IP intranet, and you can take advantage of the RSVP programming
interfaces now becoming more and more common, you can design your own
network and applications to take advantage of prioritizing. Doing so can help
you avoid buying more high cost bandwidth.
RSVP is supported by IBM OS/2 TCP/IP Version 4.1 or later, and can run as a
standalone service, or under the INETD super server. For more information
on RSVP, please consult the Online Information folder in OS/2 TCP/IP or
IETF RFC 2205.
10.5.3 Communications server
Although TCP/IP has achieved tremendous popularity, it isn’t necessarily the
best protocol for all purposes. Other network protocols have particular
strengths and weaknesses. You can take advantage of the strengths of other
protocols while still preserving the universal reach of TCP/IP on your network.
For example, IBM SNA protocols work particularly well over WAN
connections. If your network consists of remote offices, perhaps linked
together with low bandwidth connections, then you may be able to improve
the performance of TCP/IP applications by encapsulating that traffic in SNA
protocols across the WAN. Benchmark results indicate that you can get more
TCP/IP traffic across low bandwidth LAN connections if it’s wrapped into SNA
at one end and unwrapped at the other, particularly if you take advantage of
SNA’s end-to-end data compression.
The IBM SecureWay Communications Server, available for a variety of
platforms from PCs (OS/2 Warp, Windows NT) to mainframes, can be placed
at both ends of a WAN connection to encapsulate TCP/IP into SNA.
For more information on IBM SecureWay Communications Server, please
visit:
http://www.software.ibm.com/enetwork/commserver
Chapter 10. Performance
409
410
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Chapter 11. Managing your network
This chapter provides insight into how to respond to changes in your network
quickly and easily. Read how to make changes to your TCP/IP network from
any location, adding or deleting new users and connections as needed. Plus,
we'll reveal how to generate reports that show how your network is being
used so you can keep close tabs on its performance and security.
11.1 Remote administration of DHCP and DDNS
Remote administration of the DHCP and DDNS servers can be accomplished
using more than one method, and from more than one type of client platform.
Since the administration programs are Java applications, these programs can
be run remotely with a little bit of extra work.
You have two options for remote administration:
• One method of remote administration can be accomplished without adding
any extra software beyond what is delivered with OS/2 Warp Server for
e-business and OS/2 Warp Version 4 (TCP/IP Version 4.21 is either
included or can be installed). This method runs the Java-based
administration programs remotely.
• Another method adds a Web server to the OS/2 Warp Server for
e-business platform. We used Domino Go Webserver, but any Web server
that supports mapping should work. Domino Go Webserver is available for
free download. If you want to use the Secure Sockets Layer Security, you
will have to register the product, but this is not mandatory.
We will describe both methods of setting up remote administration, as well as
any problems we encountered in this process.
There are products available that allow you to administer TCP/IP services on
your company’s network across the world. QIP, from Lucent Technologies, is
one of these products that allows the management and administration of
TCP/IP services on multiple platforms, including UNIX and Windows NT. The
QIP demo software is available from
http://www.lucent.com/ins/products/qip.html.
Before going any further, an explanation is in order. The administration
programs for both OS/2 DHCP and DDNS are supplied as Java programs.
The programs are compiled so that they can run as stand-alone programs or
as applets from within a Java-enabled browser, or using an applet viewer.
© Copyright IBM Corp. 1998, 2000
411
11.1.1 No extra software
Without a Web server on the OS/2 Warp Server for e-business platform, you
can still remotely administer the DHCP/DDNS server by invoking similar
scripts to those invoked locally on the server. The difference in the scripts is
that the paths used have to reflect the network drive as seen from the client.
We have included some modified command files that will run on OS/2 or
Windows 95/98/NT with Java and REXX. REXX is shipped with OS/2 and is
available for Windows platforms. You can also follow the instructions included
with the TCP/IP product:
1. In the TCP/IP Information folder, double-click on TCP/IP Guide.
2. Search for JAVA and choose Installing the TCP/IP JAVA Configuration
Applications on a Remote Workstation.
We tested those instructions on OS/2 Warp Version 4 and Windows 95, and
they worked well for us. The command files that run on the server (and are
mentioned in the online documentation) are as follows. Our rewritten scripts
add a capital R to the front of the script names and they allow you to specify
the server you would like to administer. Note that in the case of
DHCPSCPC.CMD, the new name is RDHCPSCP.CMD (the final C is dropped
in the name to keep an 8.3 file name).
• DHCP Server Administration
• DHCPSCPC.CMD
It represents the main GUI to configure the DHCP server.
• DHCP Server Administration
• DADMGUI.CMD
It shows DHCP server statistics and provides a function to reinitialize
the DHCP server.
• DDNS Server Administration
• DDNSAPC.CMD
It represents the actual DDNS server configuration.
Here are the rewritten scripts that allow remote administration from an OS/2
workstation using Netscape 4.61 and Java 1.1.8. If you are using Netscape
4.61with Java 1.1.6 or later, you don’t need to turn off the Just In Time
compiler (JIT), which can be done by setting the JAVA_COMPILER environment
variable to any value. See 11.1.2, “Using a Web server” on page 418 for more
information on Java-version-specific dependencies.
412
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Note
If installing Java Version 1.1.8 for OS/2, do not forget to extract the original
install file using the -di and -ov parameters.
The server drive that contains the following scripts would get mapped as a
network drive for the client. The client will locally execute the rewritten scripts
on that network drive.
Note: Lines that are longer than 80 characters, such as SET CLASSPATH, are
wrapped to the next line.
RDDNSAPC.CMD is the rewritten DDNSAPC.CMD.
Chapter 11. Managing your network
413
/*
** file
: Rddnsapc.cmd
**
Remote DHCP Server Configuration
** purpose: to invoke DHCP Server Configuration from a client machine that
**
has a network drive mapped to the TCPIP drive of the server.
**
This script must be located on the server drive with \tcpip installed.
**
This script will automatically pick off the drive letter and
**
use it correctly.
**
** CAVEATS:
** 1) Client machine must have REXX installed. REXX is available for
**
Windows 95/NT also (see http://rexx.hursley.ibm.com/rexx/rexxibm.htm)
**
** history:
** 02 May
1998 added command line argument for server name to configure, m3m
** 27 April 1998
** modified from ddnsapc.cmd by Michael McDaniel (m3m), The Fourth Crusade
**
**
*/
call RxFuncAdd ’SysLoadFuncs’, ’RexxUtil’, ’SysLoadFuncs’
call SysLoadFuncs
/* trace ?i */
parse arg server
if server == ’’ then
do
parse source os . cmd
say
say ’usage:’ cmd ’servername to configure’
say
rc = beep(750,750);
EXIT
end
parse source os . cmd
JD = substr(cmd,1,1)
/* JD for Java Drive */
signal on syntax name badsyntax
signal on halt name halt
original_dir = directory();
rc = directory(JD’:\tcpip\java’);
’SET JAVA_COMPILER=XYZ;’
’SET
CLASSPATH=’JD’:\TCPIP\java\ddnssgui.jar;’JD’:\TCPIP\java\tcpauth.jar;’JD’:\TCPIP\jav
a\netdiver.jar;’JD’:\TCPIP\java\jvc.jar;’JD’:\TCPIP\java\%tcplang%\ddnsres.jar;’JD’:
\TCPIP\java\%tcplang%\tcpares.jar;’
’start /c/min java COM.ibm.raleigh.ddnssgui.client.DDNSAdministratorClient
%tcplang%’ server
rc = directory(original_dir);
EXIT
Figure 271. RDDNSAPC.CMD (Part 1 of 2)
414
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/*********************************************************************/
halt:
rc = directory(original_dir);
say 'halt occurred at line' sigl
EXIT
badsyntax:
say 'bad syntax at line' sigl
rc = directory(original_dir);
return rc
/*********************************************************************/
/* end Rddnsapc.cmd */
Figure 272. RDDNSAPC.CMD (Part 2 of 2)
RDADMGUI.CMD is the rewritten DADMGUI.CMD.
/*
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
*/
file
: Rdadmgui.cmd
Remote DHCP Server Administration
purpose: to invoke DHCP Server Administration from a client machine that
has a network drive mapped to the TCPIP drive of the server.
This script must be located on the server drive with \tcpip installed.
This script will automatically pick off the drive letter and
use it correctly.
CAVEATS:
1) Client machine must have REXX installed. REXX is available for
Windows 95/NT also (see http://rexx.hursley.ibm.com/rexx/rexxibm.htm
history:
27 April 1998
modified from dadmgui.cmd by Michael McDaniel, The Fourth Crusade
call RxFuncAdd 'SysLoadFuncs', 'RexxUtil', 'SysLoadFuncs'
call SysLoadFuncs
/* trace ?i */
parse source os . cmd
JD = substr(cmd,1,1)
/* JD for Java Drive */
signal on syntax name badsyntax
signal on halt name halt
original_dir = directory(JD':\tcpip\java');
Figure 273. RDADMGUI.CMD (Part 1 of 2)
Chapter 11. Managing your network
415
'SET JAVA_COMPILER=xxx;'
'SET
CLASSPATH='JD':\TCPIP\java\dadmgui.jar;'JD':\TCPIP\java\tcpauth.jar;'JD':\TCPIP\java
\netdiver.jar;'JD':\TCPIP\java\%tcplang%\dadmres.jar;'JD':\TCPIP\java\%tcplang%\tcpa
res.jar;'
'start /c/min java COM.ibm.raleigh.dadmgui.DHCPadmin LANG=%tcplang%'
rc = directory(original_dir);
EXIT
/*********************************************************************/
halt:
rc = directory(original_dir);
say 'halt occurred at line' sigl
EXIT
badsyntax:
say 'bad syntax at line' sigl
rc = directory(original_dir);
return rc
/*********************************************************************/
/* end Rdadmgui.cmd */
Figure 274. RDADMGUI.CMD (Part 2 of 2)
RDHCPSCP.CMD is the rewritten DHCPSCPC.CMD file.
416
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
/*
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
**
*/
file
: Rdhcpscp.cmd
Remote DDNS Server Configuration
purpose: to invoke DHCP Server Configuration from a client machine that
has a network drive mapped to the TCPIP drive of the server.
This script must be located on the server drive with \tcpip installed.
This script will automatically pick off the drive letter and
use it correctly.
CAVEATS:
1) Client machine must have REXX installed. REXX is available for
Windows 95/NT also (see http://rexx.hursley.ibm.com/rexx/rexxibm.htm
history:
02 May
1998 added command line argument for server name to configure, m3m
27 April 1998
modified from dhcpscpc.cmd by Michael McDaniel, The Fourth Crusade
^^^^^^^^
note the last 'c' was dropped to keep 8 char file name
call RxFuncAdd 'SysLoadFuncs', 'RexxUtil', 'SysLoadFuncs'
call SysLoadFuncs
/* trace ?i */
parse arg server
if server == '' then
do
parse source os . cmd
say
say 'usage:' cmd 'servername to configure'
say
rc = beep(750,750);
EXIT
end
parse source os . cmd
JD = substr(cmd,1,1)
/* JD for Java Drive */
signal on syntax name badsyntax
signal on halt name halt
original_dir = directory();
rc = directory(JD':\tcpip\java');
'SET JAVA_COMPILER=XYZ;'
'SET
CLASSPATH='JD':\TCPIP\java\dhcpsgui.jar;'JD':\TCPIP\java\ddnssgui.jar;'JD':\TCPIP\ja
va\tcpauth.jar;'JD':\TCPIP\java\netdiver.jar;'JD':\TCPIP\java\jvc.jar;'JD':\TCPIP\ja
va\%tcplang%\dhcpres.jar;'JD':\TCPIP\java\%tcplang%\tcpares.jar;'
Figure 275. RDHCPSCP.CMD (Part 1 of 2)
Chapter 11. Managing your network
417
'start /c/min java COM.ibm.raleigh.dhcpsgui.view.DHCPServerGUI %tcplang%' serverrc =
directory(original_dir);
EXIT
/*********************************************************************/
halt:
rc = directory(original_dir);
say 'halt occurred at line' sigl
EXIT
badsyntax:
say 'bad syntax at line' sigl
rc = directory(original_dir);
return rc
/*********************************************************************/
/* end Rdhcpscp.cmd */
Figure 276. RDHCPSCF.CMD (Part 2 of 2)
11.1.2 Using a Web server
The previous section described the Java programs for administration. We can
also accomplish the same tasks by using a browser to access the Java-based
administration programs. We describe the administration of OS/2 DHCP and
DDNS using, as an example, Lotus Domino Go Webserver Version 4.6.2.6
from OS/2 Warp Server for e-business. You can also use Domino Go
Webserver 4.6.2.5, which is available from the following site:
http://www.software.ibm.com/webservers/dgw/prodlist.htm
11.1.2.1 Web browser considerations
We remotely administered our OS/2 DHCP and DDNS servers successfully
from multiple platforms using multiple tools. We used Java 1.1.x-enabled
browsers and the applet viewer included with Java 1.1.8 for OS/2, also
included with Sun JDK 1.1.5 for Windows 95, NT, and Sun JDK 1.1.3 on a
Sun UltraSparc workstation.
Note: On the OS/2 platform, we recommend using at least Java 1.1.8,
available from the IBM Software Choice Web site, rather than previous
versions.
We used Sun’s HotJava browser on each of these platforms, as well to
successfully remotely administer our servers. We used Netscape V4.61 on
OS/2 and Netscape V4.7 on Windows 95, Windows 98 and Windows NT.
418
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
DHCP Server Administration
OS/2 Warp Server for e-business DHCP Server Administration does not
function properly with Netscape Communicator. Communicator uses its
own standard classes, and not those of the installed Java version. We
recommend that you use Internet Explorer for DHCP Server Administration.
Note that DDNS Server Administration works fine from Netscape
Communicator.
This section tells you how to set up the server to enable remote management
using either a browser or the applet viewer included with Sun’s JDKs.
Always remember to download the latest browser code available and to use
the latest JDKs available (assuming, of course, that you are not back-level for
some other specific reason).
At the time this publication was written, the following Netscape for OS/2
browser configurations were possible:
• Netscape V4.61 (the latest revision from Software Choice) with Java 1.02
(The Java version that comes with OS/2 Warp 4).
This configuration does not support remote configuration of TCP/IP
Version 4.1 or later for OS/2.
• Netscape V4.61 (the latest revision from Software Choice) with Java 1.1.4
(along with the latest fixes for Java 1.1.4).
We recommend turning off the Just In Time (JIT) compiler. You can do this
by setting the environment variable, JAVA_COMPILER, to a nonsense
value such as XYZ either in the CONFIG.SYS file, or temporarily from an
OS/2 command line by typing the following commands:
SET JAVA_COMPILER=XYZ
NETSCAPE
This ensures Netscape will recognize that the JIT is turned off. If using
Java 1.1.4 for OS/2, you can obtain the latest fixes from the following FTP
site:
ftp://ftp.hursley.ibm.com/pub/java/fixes/os2/11/114
• Netscape V4.61 (the latest revision from Software Choice) with Java 1.1.8.
We recommend this configuration because it ensures the best results. You
don’t need to turn off the Just In Time compiler.
By default, the following two scripts will be running when you start your
DDNS/DHCP server:
Chapter 11. Managing your network
419
DDNSAPS.CMD
DHCPSCPS.CMD
Both files reside in the \TCPIP\BIN directory. These are the scripts that allow
remote configuration. Actually, if you can locally start the DDNS and DHCP
configuration, then these two scripts are already running. When the
respective script for the DDNS or DHCP servers is not running and you try
(locally) to start the configuration program, you receive the following error:
Communication could not be established with the server.
11.1.2.2 Domino Go Webserver for OS/2
The installation of Domino Go Webserver is not described here, but it is
straightforward and we selected all options during the installation. Make sure
to write down the user ID and password you select to define during the install.
The configuration file for the Domino Go Webserver is
\MPTN\ETC\HTTPD.CNF, and we suggest that you back up the original file
before doing any configuration. You can modify the file directly, or you can
start your browser and use the GUI to configure Domino Go Webserver.
Using the browser has the advantage of proving that the Domino Go
Webserver is also running. At the server, enter the following Webserver URL:
http://localhost
You should get the Configuration and Administration Forms page. If not, enter
the following Webserver URL to see if that works:
http://localhost/admin-bin/cfgin/initial
If you are remote, replace localhost by the fully qualified domain name
(FQDN) of the HTTP server.
You will be prompted to enter user ID and password you defined when you
installed the Domino Go Webserver.
Presumably, you have the Configuration and Administration page now. Here
is how to configure Domino Go Webserver for DDNS and DHCP Remote
Administration.
In our example, the following URL brings you to the Configuration and
Administration Forms:
http://merlot.armonk.cooking.net/admin
420
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Where merlot.armonk.cooking.net represents the fully qualified domain name.
To configure Domino Go Webserver to support remote configuration of
TCP/IP Version 4.21 for OS/2, perform the following steps:
1. Make sure you are on the Configuration and Administration Forms
page, as shown in Figure 277.
Figure 277. [Go Webserver] Configuration and Administration Forms
2. Scroll down to Request processing.
3. Click on Request Routing to display the Request Routing page as shown
in Figure 278 on page 422.
Chapter 11. Managing your network
421
Figure 278. [Go Webserver] Request Routing page
4. As shown in Figure 279 on page 423, select the index number at the
bottom of the list (probably 15), select the Insert before radio button, and
then change the Action to Pass. In the URL request template field, enter
/TCPIP/*. In the Replacement file path field, enter C:\TCPIP\JAVA\*,
assuming C: is the drive where TCP/IP is installed on the server.You may
also add the Server IP Address or Host Name in the respective field.
422
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 279. [Go Webserver] Insert Routing Request
5. Click on Apply, then when the Confirmation page appears, click on
Restart Server. You should get a Restart Confirmation screen at this
point.
The following entry was inserted in the Routing Request table:
Table 15. Routing request table entry
Index
Action
Request Template
Replacement File Path
15
Pass
/TCPIP/*
C:\TCPIP\JAVA\*
6. On your server, in the C:\TCPIP\JAVA\EN_US directory, copy the
TCPLOGIN.HTM file to WELCOME.HTM.
7. You may now test the setup by executing the following command from an
OS/2 command prompt:
appletviewer http://localhost/tcpip/en_us/ddnsap.htm
Note: If not executed at the server, localhost should be replaced by the
fully qualified domain name.
Chapter 11. Managing your network
423
Alternatively, as shown in Figure 280, use the OS/2 Netscape browser with
Java 1.1.8 for OS/2 and load the following URL:
http://merlot.armonk.cooking.net/tcpip/en_us/welcome.htm
Figure 280. [OS/2 Warp] Remote Configuration of TCP/IP Version 4.21 for OS/2
Where merlot is the hostname and armonk.cooking.net is the domain name
where Domino Go Webserver is installed.
A very reliable way we have found to run the remote applets is using an
applet viewer. The applet viewer is included with the Sun Java Development
Kits for various platforms. The invocation line from a client follows for each
portion of the management. Note that we copy our appletviewer binary file
over to an applet binary file so we can type less. On OS/2 and Microsoft
platforms, just copy APPLETVIEWER.EXE to APPLET.EXE in the JDK bin
directory.
• DHCP Server Configuration
•APPLET http://<hostname>.<domain>/tcpip/java/en_us/dhcpscp.htm
This command loads the main GUI to configure the DHCP server.
424
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• DHCP Server Administration
•APPLET http://<hostname>.<domain>/tcpip/java/en_us/dadmgui.htm
This command shows server statistics and may also be used to
reinitialize server.
• DDNS Server Administration
•APPLET http://<hostname>.<domain>/tcpip/java/en_us/ddnsap.htm
This command loads the DDNS server configuration program.
For browser-based management, we used Sun’s HotJava Version 1.1.2 and
Version 3.0 browser. We tested the HotJava browser on OS/2 Warp 4 using
JDK 1.1.4 (including fixes) and also on Windows 95 and NT using Sun’s
JDK1.1.5 for Windows 95/NT. The other browsers were tested on OS/2 and
Windows 95, 98, NT and 2000.
The good news is that once the Java management applets download to your
workstation, the management application can run very fast. The management
applets are quite large, though, and take some time to download.
Unfortunately, the HotJava browser does not provide feedback as to how far
into the download you are; so patience is necessary. Once the applets
download, they communicate directly with applications on the server and
bypass the Web server. The applications themselves do not have a busy
indication; so, you will have to wait while the Java client applications are
communicating with the server applications.
The times you can expect for applet downloads are included below. It did not
seem to matter whether the applet was invoked from a browser or using the
applet viewer application; the download time was comparable for the same
platform. Download speeds were comparable for the various platforms.
Remember that these times were achieved on a 16 Mb/second token-ring
network. Scale your own times accordingly.
dhcpscp.htm - Seventeen seconds to get to the Start button, 45 seconds to
get the initial notebook after you enter your password
dadmgui.htm - Eleven seconds to get to the Start button, two seconds to get
to the main screen after you enter your password
ddnsap.htm - Twenty-four seconds to get to the Start button, 15 seconds to
get to the initial notebook screen after you enter your password
Chapter 11. Managing your network
425
11.1.2.3 Lotus Domino
If you want to do remote administration from a Lotus Domino server, it is just
as easy as from a Domino Go Webserver. You can have a Lotus Domino
server on the same machine as your DHCP/DDNS server or on a different
machine. If you have the option, we suggest you run your Domino server on a
different machine for performance reasons. If you set it up like this, you will
still need a Web server on the DHCP/DDNS server to serve up the HTML
pages and Java applets. The Domino page would just be redirected to the
other Web server. This configuration might occur if you already had an
externally accessible Domino server and you wanted to add remote
administration for a new DHCP/DDNS server that you were configuring. In
either case, the Domino configuration is similar, and we will cover both cases.
We shall start with a newly installed Domino server and explain how to set up
the needed Domino database. If you already have a Web administration
database configured, you may skip a few steps.
1. Start Domino Server Administration.
2. Click on [File — Database — New].
The server is Local and the title does not matter (we suggest Domino Web
Server Configuration), but the File Name must be DOMCFG.NSF, and
highlight Domino Web Server Configuration for your template in the lower
half of the screen. Note that the file name will not be correct unless you
type it in yourself.
3. Click OK.
4. The next page that will come up is an About Domino Web Server
Configuration. After reading it, just press Escape to get to the actual
configuration page. You may want to slide the vertical bar, a little to the
right so that you can read the configuration choices easily.
5. If your DHCP/DDNS servers are on the same machine, you will map your
URL to a directory. If your DHCP/DDNS servers are on a different machine
with a Web server installed, you will redirect your URL to another URL.
Follow steps 4 to 5 from 11.1.2.2 on page 420, or steps 9 to 10 from this
section, according to your needs. Both sets of instructions start at the
Domino Web Server Configuration page, which you can get to from the
start up configuration page by selecting [File => Database => Open] and
choosing the database.
6. Select [Create => Mapping URL=> Directory]. Skip the IP Address; fill in
your chosen comment and use the following information:
URL path: /TCPIP/*
Directory: C:\TCPIP\JAVA
426
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
assuming TCP/IP is installed on the C: drive. Check that the Read radio
button is set for Access.
7. Select [File => Save] and then [File => Close] to get back to the main
administration screen.
8. Restart your Lotus Domino HTTP server (enter HTTP quit, load HTTP) or
start Domino, as the case may be. Test by entering the following address
in a Web browser:
http://server.domain/tcpip/en_US/tcplogin.htm
You should see the TCP/IP Administration page.
9. Select [Create => Redirection URL => URL]
10.Skip the IP Address; add your comment and fill in the following:
URL path: http://thisserver.domain/tcpadmin
Redirection URL: http://otherserver.domain/tcpip/en_US/tcplogin.htm
This presumes you set up the other DHCP/DDNS Web server as
described in Part 11.1.2, “Using a Web server” on page 418.
Chapter 11. Managing your network
427
428
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Appendix A. Creating DHCP boot diskettes
This appendix examines how to create bootable DOS and OS/2 Warp
diskettes which work with your dynamic TCP/IP network, so you can install
Windows, or another kind of software, across the network to new PCs.
A.1 OS/2 Warp boot diskettes
Using three diskettes, you can start almost any PC with OS/2 Warp 4, attach
it to the network using either a TCP/IP file system or FTP, and start copying
files to install OS/2 Warp 4 or other OS/2 Warp software to a new PC. With
this technique, you can quickly and easily partition, reformat, and load hard
disks with OS/2 Warp software across a Dynamic IP network.
This section describes how to create a set of boot diskettes to attach to the
network using DHCP. The major challenge with OS/2 Warp 4 is to find a way
to fit enough advanced TCP/IP networking software on just three diskettes.
Using some compression tricks, a virtual disk, and a bit of ingenuity, you can
design your own set to work with just about any PC while still preserving the
advantages of a Dynamic IP network where addresses are available
on-the-fly.
A.1.1 Prerequisites
Before starting, you should have the following resources at your disposal:
• Four blank, formatted diskettes (1.44 MB 3.5 inch).
• Original IBM OS/2 Warp 4 CD-ROM.
• Installed OS/2 Warp 4 system with CD-ROM drive and diskette drive,
where you can prepare the boot diskettes.
• Info-Zip’s Zip and UnZip utilities, or suitable equivalents, available on the
Internet at:
ftp://ftp-os2.nmsu.edu/pub/os2/util/archiver
• If you are planning to use FTP to transfer files, a simple OS/2 Warp FTP
utility such as GET105.ZIP, written by Keith Cotroneo, is available on the
Internet at:
ftp://ftp-os2.nmsu.edu/pub/os2/apps/internet/ftp/client
• If not planning to use FTP, use another file system, such as the IBM NFS
Kit for OS/2 Warp TCP/IP (IBM Part No. 65G1255) or SRVIFS-IP by Sam
Detweiler (included on the CD-ROM accompanying this book).
© Copyright IBM Corp. 1998, 2000
429
• Server(s) able to provide appropriate file system and Dynamic IP
capabilities on the network, such as an FTP server.
• A client PC attached to the network and available for testing.
• If not provided with OS/2 Warp 4, driver files for the client network adapter,
including a correct NIF file.
A.1.2 Step-by-step procedure
We assume that you have the OS/2 Warp 4 CD-ROM available in Drive X. We
also assume that Drive A contains the diskettes you will be working on.
Substitute the correct drive letters in the following steps if necessary.
1. Create basic OS/2 Warp 4 bootable diskettes. A utility called SEDISK is
provided on the OS/2 Warp 4 CD-ROM to help prepare diskettes
appropriate to the particular type of system you plan to use. From the
OS/2 Warp command line, type:
X:\CID\EXE\OS2\SEDISK /S:X:\OS2IMAGE /T:A: /P:36
The last part (/P:36) is optional. Our test system was an IBM ThinkPad
760ED, and /P:36 adds PCMCIA support to the diskettes for this particular
notebook computer. The text file X:\SAMPLE.RSP lists other numeric
codes available for many other notebook computers.
SEDISK will require three of your four diskettes. Label the first Installation
Diskette (Diskette 0), the second Diskette 1, and the third Diskette 2. Set
these diskettes aside.
2. Create a diskette with basic TCP/IP networking. A utility called THINLAPS
is provided on the OS/2 Warp 4 CD-ROM to help add files necessary to
attach to a network. From the OS/2 Warp command line, type:
X:\CID\IMG\MPTS\THINLAPS X:\CID\IMG\MPTS A: IBMTOKCS.NIF /TCPIP /DHCP
IBMTOKCS.NIF is the name of the NIF file for the IBM PCMCIA
Token-Ring Adapter, the adapter installed in our ThinkPad 760ED. You
should substitute the name of the NIF file for your network adapter. (The
NIF files are installed in the \IBMCOM\MACS directory on your OS/2 Warp
4 system. You can view these with any text editor if you wish. Consult the
online Network Adapters and Protocol Services Guide if you wish to use a
NIF file and driver that does not ship with OS/2 Warp 4.)
THINLAPS will ask you to reinsert Diskette 1 in order to modify
CONFIG.SYS as required. Label the fourth diskette Diskette 3 - Network
Files and set it aside.
3. Edit CONFIG.SYS on Diskette 1. Use TEDIT to make some changes to
CONFIG.SYS. (You may wish to save a copy of the original in case you
430
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
make a mistake.) Our CONFIG.SYS for the ThinkPad 760ED follows, and
changes are highlighted.
buffers=32
iopl=yes
memman=swap,delayswap
protshell=CMD.EXE /K A:\STARTUP.CMD
set os2_shell=cmd.exe
diskcache=64
protectonly=yes
libpath=.;\;
ifs=hpfs.ifs /c:64
pauseonerror=no
codepage=850
devinfo=kbd,us,keyboard.dcp
devinfo=scr,ega,vtbl850.dcp
device=\dos.sys
REM device=\mouse.sys
set path=\;Z:\CMD;Z:\EXE;Z:\DLL;
set dpath=\;Z:\CMD;Z:\EXE;
set keys=on
basedev=ibmkbd.sys
basedev=ibm1flpy.add
basedev=ibm1s506.add
basedev=ibm2flpy.add
basedev=ibm2adsk.add
basedev=ibm2scsi.add
basedev=ibmint13.i13
basedev=os2dasd.dmd
device=\testcfg.sys
basedev=xdfloppy.flt
REM device=\refpart.sys
REM device=\pmdd.sys
REM set video_devices=vio_vga
REM set vio_vga=device(bvhvga)
REM set copyfromfloppy=1
BASEDEV=PCMCIA.SYS
BASEDEV=IBM2SS04.SYS
BASEDEV=IBM2SS14.SYS
DEVICE=\VDISK.SYS 2000,,
SET SAVECONNECT=1
rem *** Start of ThinLAPS additions ***
call=netbind.exe
run=lanmsgex.exe
device=lanmsgdd.os2
device=protman.os2 /I:a:\
device=sockets.sys
Appendix A. Creating DHCP boot diskettes
431
device=afinet.sys
device=ifndis.sys
device=IBMTOKCS.OS2
rem *** End of ThinLAPS additions ***
IFS=SRVIFS2.IFS
SET DIRCMD=/O:GN /N /V
Changes to CONFIG.SYS include:
• PROTSHELL was changed to run STARTUP.CMD, a batch file, used to
copy files, attach to the network, and automatically start the tasks that
you wish to perform.
• DISKCACHE, the disk cache for the FAT file system, was lowered to 64
to conserve memory. You can adjust the FAT and HPFS disk caches
according to your particular performance needs and memory
constraints.
• Certain directories in LIBPATH were removed to prevent potential
access to DLLs in unknown locations.
• DEVINFO was changed to EGA, primarily to conserve disk space.
Since only a text display is available from a diskette boot, and since all
modern PCs are still compatible with EGA, it makes little sense to
provide space-consuming VGA display services.
• MOUSE.SYS is not needed since mouse support is not required with a
simple text screen.
• PATH and DPATH have been modified to provide access to Drive Z, a
TCP/IP network drive provided by NFS or SRVIFS-IP. Feel free to
customize these lines depending on your needs.
• REFPART.SYS and PMDD.SYS device drivers are omitted, again to
conserve disk space. The services provided by these device drivers are
not required by most systems.
• VGA support statements have been REMmed out as well, again to
conserve disk space.
• SET COPYFROMFLOPPY is REMmed out to prevent OS/2 Warp’s
installation program from copying files from these diskettes.
• PCMCIA.SYS, IBM2SS04.SYS, and IBM2SS14.SYS are loaded to
provide PCMCIA support for our ThinkPad 760ED. These lines may be
different in your CONFIG.SYS (or may not exist at all for a desktop PC).
We removed all parameters from these lines to avoid some error
messages.
• VDISK.SYS is loaded to provide a 2 MB (2000) virtual RAM disk. This
RAM disk provides space for loading DLLs, programs, and other files
432
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
from compressed form on the diskette. It’s needed because we cannot
expect hard disk or network drive space during the boot process. Note
that VDISK.SYS uses the file OSO001.MSG in order to display
messages during bootup. We do not have the space available on the
diskettes to hold OSO001.MSG, so an error message is displayed
during bootup. However, VDISK still loads correctly.
• Some OS/2 Warp installation routines expect SET SAVECONNECT=1
in order to provide a network connection after bootup.
• The section ThinLAPS additions in CONFIG.SYS has been modified
slightly so that the connection to the TCP/IP network does not start
right away. That process has been moved to a batch file called
SETUP.CMD, which is called from the STARTUP.CMD batch file.
IBMTOKCS.OS2 is the device driver used for the IBM PCMCIA
Token-Ring Adapter, and this line will be different for other network
adapters. Some releases of MPTS (Multiprotocol Transport Services)
include AFLEAN.SYS. AFLEAN.SYS may be used instead of
AFINET.SYS in order to save space on the diskette.
• SRVIFS2.IFS, part of SRVIFS-IP, is loaded to provide a TCP/IP file
system. You may elect to use NFS instead or omit this line entirely if
using FTP.
• The SET DIRCMD line is entirely optional. It’s used here so that a
directory listing (DIR) is sorted alphabetically and formatted differently.
4. Delete unnecessary files on Diskette 2 and add necessary files from
Diskette 3. Diskette 2 (the third diskette) is the last diskette actually used
in the boot process. All the required networking files from Diskette 3,
created by THINLAPS, need to be squeezed onto Diskette 2.
In order to properly merge these diskettes, try copying all the files on
Diskettes 2 and 3 to a temporary directory on your hard disk. Delete,
modify, and move the files in this working directory as needed. Here’s a
listing of the files that should be on Diskette 2 when you’re done:
AFINET
BKSCALLS
BMSCALLS
BVHINIT
BVSCALLS
CMD
COUNTRY
DOS
DOSCALL1
FILES
HPFS
SYS
DLL
DLL
DLL
DLL
EXE
SYS
SYS
DLL
EXE
IFS
225479
8-11-96
512
8-12-96
512
8-12-96
10540
8-12-96
512
8-12-96
74640
8-09-96
36185
8-09-96
1142 12-04-95
123680
8-27-96
482698
4-03-98
141378
8-13-96
10:05p
3:00a
3:02a
2:08a
2:59a
12:15a
12:29a
11:22p
8:45a
11:22a
11:17a
Appendix A. Creating DHCP boot diskettes
433
IBMTOKCS
IFNDIS
KBDCALLS
KEYBOARD
LANMSGDD
LANMSGDL
LANMSGEX
LT0
LTG
MSG
NAMPIPES
NETBIND
NLS
OS2CHAR
PRO
PROTMAN
PROTOCOL
QUECALLS
SESMGR
SETUP
SOCKETS
SRVIFS2
STARTUP
TESTCFG
VDISK
VIOCALLS
VTBL850
OS2
SYS
DLL
DCP
OS2
DLL
EXE
MSG
MSG
DLL
DLL
EXE
DLL
DLL
MSG
OS2
INI
DLL
DLL
CMD
SYS
IFS
CMD
SYS
SYS
DLL
DCP
28724
35828
1024
28097
3604
2580
1099
12943
4487
512
1024
13657
512
512
2234
22308
412
1024
1536
122
59565
23087
1805
9808
2904
2048
10478
4-30-96
8-11-96
8-12-96
8-09-96
8-01-96
8-01-96
8-01-96
8-01-96
3-28-96
8-12-96
8-12-96
8-01-96
8-12-96
8-12-96
8-01-96
8-01-96
4-02-98
8-12-96
8-12-96
4-03-98
8-15-96
1-06-93
4-03-98
8-12-96
8-12-96
8-14-96
8-09-96
2:06p
10:04p
3:02a
12:58a
6:14p
6:14p
6:14p
6:14p
10:11a
2:56a
3:02a
6:13p
2:55a
2:58a
6:13p
6:13p
5:15p
2:57a
2:52a
10:38a
3:47p
9:39p
10:13a
2:21a
2:24a
5:06a
1:00a
Some comments on particular files follow:
• IBMTOKCS.OS2 is the driver for the IBM PCMCIA Token-Ring Adapter.
You may be using another driver.
• PROTOCOL.INI is a text file containing adapter and protocol settings.
In a token-ring network, for example, you may wish to edit this file to
change the default ring speed used by the IBM PCMCIA Token-Ring
Adapter. Other network adapters will have different settings.
THINLAPS attempts to create a default PROTOCOL.INI which works in
most situations.
• SRVIFS2.IFS is the file system driver for SRVIFS-IP. If you are not
using SRVIFS-IP, this file is not required.
• VDISK.SYS can be found in the \OS2\BOOT directory on your OS/2
Warp 4 system.
FILES.EXE, SETUP.CMD, and STARTUP.CMD require some additional
work to set up properly.
434
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
FILES.EXE is a self-extracting, compressed Zip file, containing the files
needed to attach to the network. (We created FILES.EXE using Info-Zip’s
Zip and Zip2Exe utilities.) These files are extracted to the RAM disk
created by VDISK.SYS during the boot process. Many of the remaining
files on Diskettes 2 and 3 are contained in FILES.EXE, but not all. For our
ThinkPad 760ED, we included the following files in FILES.EXE:
Archive:
Length
-----512
19893
11587
168538
24710
0
13218
5451
42725
55296
9415
17193
24064
21812
35552
20671
37842
9035
21947
62289
47279
34304
32288
9390
14596
-----739607
FILES.EXE
Size
Ratio
-------205
60%
11872
40%
6346
45%
131892
22%
19384
22%
0
0%
3619
73%
2079
62%
9073
79%
27572
50%
7731
18%
10234
41%
13180
45%
16318
25%
19216
46%
12002
42%
15905
58%
6266
31%
12813
42%
30200
52%
24151
49%
9344
73%
17551
46%
8330
11%
3053
79%
-------418336
43%
Date
---08-12-96
08-23-96
08-23-96
08-11-96
08-11-96
04-02-98
08-11-96
08-11-96
08-11-96
01-10-98
08-13-96
08-23-96
08-23-96
08-12-96
08-23-96
08-23-96
08-23-96
01-31-93
06-07-93
08-23-96
08-23-96
08-11-96
08-23-96
11-04-96
11-04-96
Name
---ANSICALL.DLL
ARP.EXE
CNTRL.EXE
DHCPCD.EXE
DHCPSTRT.EXE
ETC/
ETC/DHCPCD.CFG
ETC/PROTOCOL
ETC/SERVICES
GET.EXE
HARDERR.EXE
IFCONFIG.EXE
INETWAIT.EXE
NPXEMLTR.DLL
PING.EXE
ROUTE.EXE
SO32DLL.DLL
SRVATTCH.EXE
SRVRDR.EXE
TCP32DLL.DLL
TCPIPDLL.DLL
TCPMRI.DLL
TCPTIME.DLL
TEDIT.EXE
TEDIT.HLP
------25 files
Some of these files, in particular, deserve comment:
• GET.EXE is a simple FTP program we chose among the several
available on the Internet. It takes up very little space (just 27 KB
compressed) and can be used in batch files.
• PING.EXE is optional (but useful). We extracted PING.EXE from
X:\CID\IMG\MPTS\MPTN\BIN\BIN.ZIP.
• SRVATTCH.EXE and SRVRDR.EXE are part of SRVIFS-IP.
Appendix A. Creating DHCP boot diskettes
435
• TCPTIME.DLL is required by the version of INETWAIT.EXE in the
original release of TCP/IP Version 4.0 for OS/2 Warp. We extracted
TCPTIME.DLL from X:\CID\IMG\MPTS\MPTN\DLL\MDLL.ZIP.
• TEDIT.EXE and TEDIT.HLP are both optional but quite useful. This
simple text editor can be used to make changes to CONFIG.SYS,
STARTUP.CMD, and/or SETUP.CMD when troubleshooting. You can
find both these files in the \OS2 directory on your OS/2 Warp 4 system.
5. Create STARTUP.CMD and SETUP.CMD. STARTUP.CMD and
SETUP.CMD are the key batch files used at bootup. Here’s how our
STARTUP.CMD file looks:
@echo off
set vdisk=r:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=q:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=p:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=o:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=n:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=m:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=l:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=k:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=i:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=h:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=g:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=f:
436
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=e:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=d:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
set vdisk=c:
copy a:files.exe %vdisk% 1>nul 2>nul
if not errorlevel 1 goto config:
goto error
:config
%vdisk%
copy a:\*.dll
copy a:\*.exe
copy a:\*.cmd
set path=%vdisk%\;%path%
set dpath=%vdisk%\;%dpath%
set beginlibpath=%vdisk%\;a:\;
files
del files.exe
if not exist setup.cmd goto end:
call setup
goto end:
:error
Echo Error finding VDISK
:end
The first several lines in this batch file attempt to locate the RAM disk
created with VDISK.SYS, starting with Drive R and working back to Drive
C. Once located, all the necessary files from Diskette 2 are copied to the
RAM disk. Some environment variables are set (to allow FILES.EXE and
other programs to run properly); then FILES.EXE is run to extract the
compressed files inside. Finally, if SETUP.CMD exists, STARTUP.CMD
calls it.
SETUP.CMD is considerably shorter and looks like this:
SET ETC=%vdisk%\ETC
DETACH CNTRL.EXE
INETWAIT
DHCPSTRT -i lan0
DETACH SRVRDR.EXE
SRVATTCH Z: \\192.168.6.12\TESTDRV
The first line sets the ETC environment variable, required by many TCP/IP
applications. The ETC directory contains critical configuration files. Then
Appendix A. Creating DHCP boot diskettes
437
CNTRL.EXE is started as a background task. INETWAIT runs (to wait for
the TCP/IP stack to initialize); then DHCP is started (to obtain address
information and connect to the network over "lan0," in this case the IBM
PCMCIA Token-Ring Adapter).
The last two lines start SRVIFS-IP and create a network drive (Drive Z)
from 192.168.6.12, a SRVIFS-IP server. (If a name server is available, and
its address is returned by the DHCP server, the IP name can be specified
instead.) You may wish to use NFS, GET (FTP) or some other method to
obtain additional files across the network. Any additional commands can
be added to this batch file as needed.
6. Test and verify your network connection. We’ve included a few programs
on these boot diskettes so you can troubleshoot your connection to the
network. These files include:
• LANTRAN.LOG will be created on Diskette 2 and will contain
information on whether the network adapter was able to attach to the
network. Use the information in this file to determine whether the
network adapter driver is functioning correctly and whether
PROTOCOL.INI is constructed properly.
• ARP.EXE lets you look up an IP address based on a network adapter
address.
• IFCONFIG lets you determine the IP address assigned to the system.
At the command line, after booting from these diskettes, type (in
lowercase):
ifconfig lan0
• PING.EXE allows you to test the connection to a particular system. It
can be particularly helpful when you’re installing both a new network
and a new set of PCs. You can test to see whether traffic is being
routed correctly.
A.1.3 Notes on the step-by-step procedure
1. Try to use TEDIT when making changes to text files on the diskettes. The
OS/2 System Editor will create extended attributes for each text file you
edit, and these extended attributes take up extra space. Extended
attributes are not required for boot diskettes.
2. Info-Zip’s Zip utility can accept the -9 option, resulting in the highest
possible degree of compression. Then the Zip2Exe utility can convert the
Zip file to FILES.EXE. We recommend using -9 to conserve as much
space as possible. You should also use -r to properly store the files in the
ETC directory.
438
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
3. While we opted to use OS/2 Warp files obtained from the original CD-ROM
release, you may wish to use files for these boot diskettes from OS/2 Warp
FixPaks and later versions of MPTS and TCP/IP. If so, there may be
dependencies on certain DLLs that we have not included. Newer is not
necessarily better, however. Files tend to grow in size over time as
features are added.
4. SNIFFLE, part of the MPTS supplemental utilities, can be used to verify
the construction of a network adapter driver’s NIF file. See the online
Network Adapters and Protocol Services Guide for details.
A.2 DOS boot diskettes
With DOS boot diskettes that can attach to a file server, such as Warp Server
or Windows NT, you can install many different software packages across the
network, including versions of Windows with DOS-based installation
programs.
This section describes how to create a pair of DOS boot diskettes that can
provide a PC with a TCP/IP (TCPBEUI) or NetBEUI-based file system and a
dynamically assigned IP address.
A.2.1 Prerequisites
These are the necessary prerequisites:
• Two blank, formatted diskettes (1.44 MB 3.5 inch)
• A workstation that has either BASIC or FULL DOS LAN Services (DLS)
and IBM PC-DOS 2000 installed
• Server(s) able to provide appropriate file system and Dynamic IP
capabilities on the network, such as Warp Server
• A client PC attached to the network and available for testing
A.2.2 Step-by-step procedure
We will assume that you have DOS LAN Services installed on Drive C in the
\NET subdirectory. We will also assume that Drive A will contain any diskettes
you will be working with. Substitute the correct drive letters in the following
steps if necessary.
1. Create Diskette 1. Due to the size of DLS, we have to split the software
across two diskettes. Diskette 1 contains the files to boot the PC and
initialize the network adapter. It will also contain files that allow you to
partition and format a hard drive and to edit files. Label this diskette DLS
Appendix A. Creating DHCP boot diskettes
439
Boot Diskette 1. It should be formatted under IBM PC-DOS 2000 using the
FORMAT A: /S command, to create a bootable diskette.
Then copy the following files from the \DOS subdirectory to Diskette 1:
•
•
•
•
•
•
•
•
•
EMM386.EXE
HIMEM.SYS
SETVER.EXE
SHARE.EXE
FORMAT.COM
FDISK.COM
E.EXE
E.EX
E.INI
Copy the following files from the \NET subdirectory to the diskette:
•
•
•
•
•
DLSHELP.SYS
LT2.MSG
NTSTS.DOS
PROTMAN.DOS
IBMTOK.DOS
Our test system was an IBM ValuePoint with an IBM Token-Ring 16/4
network adapter, so we copied the IBMTOK.DOS driver file. You may have
to copy a different driver depending on the type of network adapter you
have in your PC.
2. Create CONFIG.SYS in order to boot a new PC and partition its hard drive
without having to wait for a network connection to be established. We can
create a menu-based CONFIG.SYS on Diskette 1 as follows:
[MENU]
MENUITEM=NETWORK, Connect to the network
MENUITEM=STANDALONE, No network connection
MENUDEFAULT=NETWORK,15
[COMMON]
FILES=30
BUFFERS=10
DOS=HIGH,UMB
DEVICE=A:\HIMEM.SYS
DEVICE=C:\DOS\EMM386.EXE noems ram x=d800-e2ff
DEVICEHIGH=C:\DOS\SETVER.EXE
[NETWORK]
DEVICEHIGH=A:\PROTMAN.DOS /i:A:\
DEVICEHIGH=A:\IBMTOK.DOS
440
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
DEVICEHIGH=A:\NTSTS.DOS
DEVICEHIGH=A:\DLSHELP.SYS
LASTDRIVE=Z
[STANDALONE]
Remember to replace the DEVICEHIGH=A:\IBMTOK.DOS statement with the
correct statement for your network adapter.
3. Create AUTOEXEC.BAT. Use the E text editor to create the following
AUTOEXEC.BAT file on Diskette 1:
SET PATH=A:\
GOTO %CONFIG%
:NETWORK
SHARE
SET TCPHELP=A:\
SET ETCDIR=A:\
ECHO ................................
ECHO Insert DLS Boot Disk 2 now, then
PAUSE
NET START
:STANDALONE
4. Create PROTOCOL.INI. The following sample PROTOCOL.INI file is from
our IBM ValuePoint. You should copy the PROTOCOL.INI file from \NET
subdirectory to Diskette 1; then use E to modify the DHCPClientID=
parameter (highlighted below):
[network.setup]
version=0x3100
netcard=ibm$genibmtok,1,IBM$GENIBMTOK
transport=nts$ntst2,NTS$NTST2
transport=ibm$netbeui,IBM$NETBEUI
lana0=ibm$genibmtok,1,nts$ntst2
lana1=ibm$genibmtok,1,ibm$netbeui
[protman]
DriverName=PROTMAN$
PRIORITY=ibm$netbeui
[IBM$GENIBMTOK]
ram=0xD800
DriverName=IBMTOK$
primary
[NTS$NTST2]
DNSAddr=
GatewayAddr=
Appendix A. Creating DHCP boot diskettes
441
NetSubNetMask=
IPAddr=
TCPHeartBeats=Standard
DHCPClientID=BURGER
BootPFlag=DHCP
DriverName=ntsts$
VCs=16
VCReceiveLarge=6
VCSends=6
RcvWindow=2920
UseMemory=UMB
BINDINGS=IBM$GENIBMTOK
LANABASE=0
Token-Ring
[IBM$NETBEUI]
DriverName=netbeui$
SESSIONS=20
NCBS=20
BINDINGS=IBM$GENIBMTOK
LANABASE=1
5. Create Diskette 2. Diskette 2 contains the files required to run DLS. Label
this diskette DLS Boot Diskette 2.
Insert Diskette 2 and copy the following files from the \NET subdirectory to
the diskette:
•
•
•
•
•
•
•
•
CMDS.EXE
CMDS16.EXE
NET.MSG
NET.EXE
NETH.MSG
NETXP.MSG
PING.EXE
PROTMAN.EXE
6. Create NETWORK.INI. The following sample NETWORK.INI file is from
our IBM ValuePoint. You should copy the NETWORK.INI file from the \NET
subdirectory to Diskette 2; then use E to modify the highlighted
parameters as needed:
[network]
computername=BURGER
lanroot=A:\
autostart=netbeui basic
username=USERID
domain=ARMONK
442
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
autologon=no
lslogon=yes
reconnect=no
passwordcaching=no
timesync=no
[install]
peer=no
gui=no
windows=no
protocol=tcpnetr
minidls=no
installed=no
target=c:~WSRCLNT\
**for minidls only
**for minidls only
**for minidls only
[Password Lists]
7. Final configuration. To make Diskette 2 work correctly, you must copy the
AUTOEXEC.BAT and COMMAND.COM files from Diskette 1. These files
should be identical on both diskettes.
By having these two files on both diskettes, you will be able to remove
Diskette 1 and replace it with Diskette 2 when the PAUSE statement in
AUTOEXEC.BAT is executed. This allows DLS to start properly.
8. Testing. Boot the computer with Diskette 1 inserted. The following screen
should be displayed:
PC DOS 2000 Startup Menu
========================
1. Connect to the network
2. No network connection
Enter a choice:
Time remaining: 15
F5=Bypass startup files F8=Confirm each line of CONFIG.SYS and AUTOEXEC.BAT [N]
Option 1 will be highlighted and will be automatically selected after 15
seconds. If selected, option 2 will exit directly to the DOS prompt.
When option 1 is selected, the network drivers will be loaded, and you will
be prompted to insert Diskette 2 and press any key to continue.
When this is done, DLS will be started and the DOS prompt will be
displayed. Note that you are not logged onto the LAN.
Appendix A. Creating DHCP boot diskettes
443
To check the IP configuration of the machine, use the PING command with
the DHCPClientID from NETWORK.INI as a parameter.
A:\>ping burger
PING - ICMP Echo Request/Reply 2.09 (960320)
Copyright (c) 1995-1996 Network Telesystems, Inc. All rights reserved.
PING burger (192.168.6.15): 56 ICMP data bytes
64 bytes from burger: icmp_seq = 0. time < 55 ms
---- burger PING Statistics ---1 packets transmitted
1 packets received
0% packet loss
round-trip (ms) min/avg/max = 55/55/55
A.2.3 Notes on the step-by-step procedure
• The test system was running IBM PC-DOS 2000.
• The test system was running the DLS code available with the IP08267
FixPak for Warp Server, which was downloaded from:
ftp://ftp.software.ibm.com/ps/products/lan/fixes/lsv5.0/english-us
Attention
The IP08267 FixPak is provided only for licensed users of OS/2
Warp Server.
444
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Appendix B. Application issues
This appendix describes some specific application issues you should be
aware of when using DHCP and/or DDNS. Generally speaking, most
applications should not have any problem running on a DHCP client machine.
There are exceptions, however. We recommend that you consult with
software vendors to make sure that your TCP/IP applications work well on a
DHCP client system. If necessary, you can configure your client machine with
a static IP address if you must run a particular application.
Many applications (such as network management utilities) can be much
easier to use with a Dynamic DNS server on the network if you do rely on
DHCP. Otherwise, users may find it difficult to locate other systems on the
network.
We have tested several popular applications with DHCP and DDNS
technologies, and, while the overwhelming majority of applications have no
problems, we have found some minor difficulties with a few specific
applications. The following sections describe those problems.
B.1 DB2 Universal Database
Many network managers prefer to assign fixed IP addresses to major servers
to prevent them from relying on other servers in order to function properly.
Quite often, DB2 servers have fixed IP addresses.
However, the client version of DB2 may be installed on a workstation which
receives its address assignment from a DHCP server. We found a problem
with DB2 UDB Version 6.1 on OS/2 Warp DHCP clients. To keep abreast of
DB2 technical information, you should visit the Web site:
http://www.software.ibm.com/data/db2/db2tech
To identify this problem, we installed the IBM DB2 Universal Database
Personal Edition on an OS/2 Warp Version 4 client. The INSTALL.TXT file
warns that the TCP/IP protocol stack will not be recognized on a DHCP client.
In fact, DB2 itself seems to work well on such a client. We could only find one
problem: The HTML online documentation system (NET.QUESTION) did not
work properly on an OS/2 DHCP client until we applied a fix to it. Here is how
we obtained the fix, and this is what it offers:
1. First, we needed to obtain the latest NetQ driver, which must not be earlier
than NQ990527.ZIP. This driver is, unfortunately, not delivered with FixPak
© Copyright IBM Corp. 1998, 2000
445
1 for DB2 UDB for OS/2 v6.1, but the latest driver may be downloaded
from:
http://cognac.boeblingen.de.ibm.com/
The driver download area of this site is only accessible to registered users,
so you must first register, and then, after receiving your user ID and
password, you may download the driver.
2. NQ990527.ZIP does not change the NETQ.CFG file, and it has
communication set to local not TCP/IP, using named pipes.
3. NETQINIT.EXE is no longer called, which means the TCP/IP NETQ.CFG file
is not created. If you require TCP/IP after the installation, then you may
call NETQINIT.EXE.
The FixPak 1 for DB2 UDB, which we referred to earlier, is available from:
http://www-4.ibm.com/software/data/db2/db2tech/version61.html
The Windows 95, Windows 98 and Windows NT versions of DB2 UDB no
longer contain the warning in INSTALL.TXT. We experienced no obvious
problems when installing or running DB2 UDB on a Windows 95, Windows 98
or Windows NT DHCP client.
B.2 TME 10 Framework
TME 10 only needs to have the correct IP addresses for the TMR, in order to
work properly. Currently, TME 10 has two ways to provide client support for
DHCP: the UserLink/DHCP Service component of PC Agent and TMA (Tivoli
Managed Agent). If possible, you should use Tivoli’s TMA, since it provides
newer technology and will gradually replace the UserLink/DHCP Service.
TMA, which many refer to as a thin client product, takes up very little disk
space.
The TMA design was intended for DHCP use, where users would shut down
their machines, reboot later (maybe even the next day) and receive a new IP
address. If an IP address changes on the TMA without warning, when a
DHCP lease expires and a new one is acquired, then problems begin. This
has been documented in APAR IX85877, and no patches have been released
yet.
Generally, the Tivoli service, a daemon, runs on each managed client and
server attached to the network. As these daemons try to connect with one
another, they must be able to map a system’s IP address to a valid hostname.
This reverse mapping technique may not be available in a DHCP-only
446
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
environment, and many ordinary DNS servers can run without reverse
mapping (for instance, no PTR record update). In fact, Tivoli products use
both hostname to IP address mapping and reverse mapping, and installation
of Tivoli clients will fail if both services are not provided on your network.
Historically, Tivoli has recommended a Microsoft Windows NT WINS server to
help map IP addresses to client computer names (at least for WINS clients
such as Windows 95, Windows 98 and Windows NT, where the host and
computer names are often identical). However, a DDNS server can provide
broader support for DHCP TMA clients.
The TMR Server must never be configured as a DHCP client.
For more information on DDNS and DHCP issues with TME 10, please refer
to the following redbooks:
• All About Tivoli Management Agents, SG24-5134
• TME 10 Internals and Problem Determination, SG24-2034
B.3 Netfinity
Netfinity Version 5.20.4 (and later), IBM PC management software for
workgroups, works well with DHCP clients. You also need a DNS (preferably
DDNS) server, and you need to keep in mind some minor configuration
issues. The following sections describe those issues.
Figure 281. [Netfinity Manager 5.20.4] Netfinity Service Manager Folder
Appendix B. Application issues
447
B.3.1 Manager
Netfinity Manager has to be explicitly configured to use dynamic addressing.
First, open the Netfinity Service Manager. Then open the Remote System
Manager as shown in Figure 281. The System Group Management folder
should appear with a single default group, All. Click on Options and then on
Dynamic Address Options as shown in Figure 282. When the Dynamic
Address Options window opens, check the checkboxes, as shown in Figure
283. (You may also wish to increase the Dynamic Ping Interval from the
default of one minute to three minutes.) Click OK to save, and you are
finished.
Figure 282. [Netfinity Manager 5.20.4] System Group Management options
Figure 283. [Netfinity Manager 5.20.4] Dynamic Address Options
B.3.2 Client (services)
The client configuration simply requires coordination between the Netfinity
system name (for TCP/IP) and the designated IP hostname for the system. As
long as these two names are the same, Netfinity Manager can track the
machine by name, even if its IP address changes.
448
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Figure 284. [Netfinity Services 5.20.4] Network Driver Configuration for a client
To configure the Netfinity client, first open the Netfinity folder and then open
the Network Driver Configuration. The Network Driver Configuration
window (Figure 284) allows you to change certain client settings. You should
make sure that the System Name is set the same as the IP hostname for the
client (as registered with the name server). Also, the TCP/IP driver should be
enabled.
You may wish to add some System Keywords to more easily identify client
systems on the network. You can assign keywords for different categories of
clients by client operating system, processor type, location, department, and
so forth.
Click on Save, and you should see a window similar to Figure 285. Click on
OK then Exit. Shutdown and reboot the system so that the changes you
made can take effect, or simply kill and restart NETFBASE.EXE.
Appendix B. Application issues
449
Figure 285. [Netfinity Services 5.20.4] Network Driver Configuration updated
You should also configure the DDNS client hostname to be identical to the
Netfinity name. How you register your client with the DDNS server will depend
on the type of client. For example, in OS/2 Warp 4, you can use the DDNS
Client Configuration program (Figure 286). Note that the hostname oyster is
identical to the Netfinity client name.
Figure 286. [OS/2 Warp Version 4] DDNS Client Configuration
450
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Once you have completed setup of your Netfinity client(s), you should be able
to monitor them using the Netfinity Manager.
B.4 Netscape Navigator
Netscape’s popular Web browser has no trouble working on any client
receiving its address assignment from a DHCP server. In fact, some versions
of Netscape Navigator can be enhanced using DHCP technologies. Please
see 7.2.3, “Netscape Message of the Day service” on page 334, and 7.2.4,
“Custom Netscape INI files” on page 341, for more information on Netscape
Navigator and DHCP.
Appendix B. Application issues
451
452
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Appendix C. Using the additional material
We have included additional utilities and information on the CD-ROM that
accompanies this redbook. You can also avoid typing in long script files or
other examples from the redbook that have been included on the CD-ROM.
C.1 Using the CD-ROM
The CD-ROM autoruns on Windows systems, starting your default browser
with the main file, INDEX.HTM. If autorun is turned off or you are not using a
Windows system, you can also browse the contents of the CD-ROM manually
by opening the \INDEX.HTM file with your favorite Web browser, such as
Netscape Communicator or Microsoft Internet Explorer. These HTML files do
not use frames, but there is a link to the HTML files of the original edition
CD-ROM, and these files did implement frames.
Figure 287 shows the main HTML page for the CD-ROM from a Netscape
browser.
Figure 287. CD-ROM Contents main page: INDEX.HTM
© Copyright IBM Corp. 1998, 2000
453
Note: The URL of file:///E|/INDEX.HTM in Figure 287 on page 453 represents
a CD-ROM drive and directory.
C.2 Locating the additional material on the Internet
Some of the information on the CD-ROM associated with this redbook is also
available in softcopy on the Internet from the IBM Redbooks Web server. You
can go to the IBM Redbooks Web site at:
http://www.redbooks.ibm.com/
Select the Additional materials button on the left part of the page, then click
to display the directory listing. Select the SG245280 directory and you will
see selected files to download.
C.3 Second edition CD-ROM contents
This redbook is the second edition, and the following content have been
added to the CD-ROM as follows:
• An RFC Library with documentation on all of the RFCs, with both summary
and detailed information, as well as a way to update the tables shown in
the pages.
• The AIX files for static DNS, DDNS and DHCP for the server and client.
• The WorkSpace On-Demand Version 2.0 Handbook in Acrobat Reader
PDF format.
• A link to the HP JetDirect Software Configuration Guide in Acrobat Reader
PDF format.
• A link to the CD-ROM contents of the original redbook, which are
described in the next section.
C.4 First edition CD-ROM contents
In addition to the content described above, there is a link to the HTML files for
the original CD-ROM. The contents of the original CD-ROM are as follows:
• A dynamic IP presentation in Lotus Freelance format
• The REXX scripts documented throughout the book
• Information to let you CID install a OS/2 Warp Server server from scratch
can be found in the \CID directory. Included is:
• A Client .CMD file to install a Warp Server server from scratch.
454
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• Response files for:
• Java 1.1.4
• OS/2 Warp Server File and Print Services (LAN Server 5)
• MPTS 5.30
• Netscape 2.02
• OS/2 Version 3.x
• TCP/IP Version 4.1 for OS/2
• IP08506 CSD OS/2 Warp Server File and Print Services
• FixPak 35 for OS/2 3.x
• Fixes and CSD’s included:
• Fixes for JAVA 1.1.4
• Fixes for MPTS 5.30
• Fixes for TCP/IP 4.1
• IP08506 CSD for OS/2 Warp Server File and Print Services (LAN
Server 5)
• IP_8508 CSD for HPFS386
• FixPak 35 for OS/2 Version 3.x
• Some useful Windows 95/NT network utilities
• AG NetTools -A collection of GUI-based tools for testing Internet and
IP-based networks.
• CyberKit - Another collection of GUI-based tools for testing Internet
and IP-based networks.
SockVer - Displays the version of Winsock you are using.
Appendix C. Using the additional material
455
456
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Appendix D. DHCP options (RFC 2132)
The Internet Engineering Task Force (IETF) publishes official documents
describing Internet standards, including DHCP. DHCP provides several
different options. This appendix, based on RFC 2132, includes a list of
options 1-67, along with information on which options are supported by
popular server and client operating systems. Options 1-127 are considered
public options and are defined as Specification Required. RFC 2489
describes options 128-254 that are site-specific options and are defined as
Private Use. RFC 2489 also documents how to go about getting an option
128-254 standardized.
DHCP provides a framework for passing configuration information to hosts on
a TCP/IP network. Configuration parameters and other control information
are carried in tagged data items that are stored in the option field of the
DHCP message. The data items themselves are also called options.
RFC 1497 was published previously and included vendor information
extensions. These extensions are also described in this appendix, and RFC
1497 should now be considered obsolete. In addition, all the DHCP options
described here, except those specific to DHCP in Part D.3, “DHCP (only)
options” on page 476, may be used as BOOTP vendor information
extensions.
For the latest, official list of DHCP options please visit:
ftp://ds.internic.net/rfc/rfc2132.txt
D.1 Introduction
DHCP options have the same format as the BOOTP vendor extensions
defined in RFC 1497. Options may be fixed length or variable length. All
options begin with a tag octet, which uniquely identifies the option. Fixed
length options without data consist of only a tag octet. Only options 0 and 255
are fixed length. All other options are variable length with a length octet
following the tag octet. The value of the length octet does not include the two
octets specifying the tag and length. The length octet is followed by (length)
octets of data. With some variable length options the length field is a constant
but must still be specified. Any options defined in a future version of this
standard must contain a length octet even if the length is fixed or zero.
Options containing NVT ASCII data should not include a trailing null.
However, the receiver of such options must be prepared to delete trailing nulls
© Copyright IBM Corp. 1998, 2000
457
if they exist. The receiver must not require that a trailing null be included in
the data.
Option codes 128 to 254 (decimal) are reserved for your own use.
D.2 DHCP and BOOTP options
This section introduces all DHCP and BootP options that can be defined at
the DHCP server and delivered to requesting DHCP and BootP clients. Be
aware that although most DHCP servers can deliver all options illustrated
here to requesting DHCP and BootP clients, there are platform-specific
dependencies that determine whether or not requesting DHCP and BootP
clients can actually make use of all delivered information.
D.2.1 Options 0 and 255: Pad and end
Originally described in RFC 1497, options 0 and 255 help provide vendor
extensions to the DHCP standard.
Option 0, the pad option, can be used to cause subsequent fields to align on
word boundaries. Its length is one octet. The end option, option 255, marks
the end of valid information in the vendor field. (Subsequent octets should be
filled with pad options.) Its length is also one octet.
D.2.2 Option 1: Subnet mask
Option 1 specifies the client's subnet mask per RFC 950. If both the subnet
mask and the router (option 3) are specified in a DHCP reply, the subnet
mask option must be first.
The code for the subnet mask option is 1, and its length is 4 octets.
Code Length
Subnet Mask
+-----+-----+-----+-----+-----+-----+
| 1 | 4 | m1 | m2 | m3 | m4 |
+-----+-----+-----+-----+-----+-----+
D.2.3 Option 2: Time offset
The time offset field specifies the offset, in seconds, of the client's subnet
from Coordinated Universal Time (UTC). The offset is expressed as a two's
complement 32-bit integer. A positive offset indicates a location east of the
zero meridian, and a negative offset indicates a location west of the zero
meridian.
458
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The code for the time offset option is 2, and its length is 4 octets.
Code Length
Time Offset
+-----+-----+-----+-----+-----+-----+
| 1 | 4 | n1 | n2 | n3 | n4 |
+-----+-----+-----+-----+-----+-----+
D.2.4 Option 3: Router
The router option specifies a list of IP addresses for routers on the client's
subnet. Routers should be listed in order of preference.
The code for the router option is 3, and the minimum length for the router
option is 4 octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 3 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.5 Option 4: Time server
The time server option specifies a list of RFC 868 time servers available to
the client. Servers should be listed in order of preference.
The code for the time server option is 4, and the minimum length for this
option is 4 octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 4 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.6 Option 5: IEN 116 (Old) name server
The name server option specifies a list of IEN 116 name servers available to
the client. Servers should be listed in order of preference. (Option 5 is
provided for compatibility with old-style name servers. Use Option 6 for
modern domain name servers.)
The code for the name server option is 5, and the minimum length for this
option is 4 octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 5 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
Appendix D. DHCP options (RFC 2132)
459
D.2.7 Option 6: Domain name server
The domain name server option specifies a list of Domain Name System
(RFC 1035) name servers available to the client. Servers should be listed in
order of preference.
The code for the domain name server option is 6, and the minimum length for
this option is 4 octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 6 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.8 Option 7: Log server
The log server option specifies a list of MIT-LCS UDP log servers available to
the client. Servers should be listed in order of preference.
The code for the log server option is 7, and the minimum length for this option
is 4 octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 7 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.9 Option 8: Cookie server
The cookie server option specifies a list of RFC 865 cookie servers available
to the client. Servers should be listed in order of preference.
The code for the log server option is 8, and the minimum length for this option
is 4 octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 8 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.10 Option 9: LPR server
The LPR server option specifies a list of RFC 1179 print servers available to
the client. Servers should be listed in order of preference.
460
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The code for the LPR server option is 9, and the minimum length for this
option is 4 octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 9 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.11 Option 10: Impress server
Option 10 specifies a list of Imagen Impress servers available to the client.
Servers should be listed in order of preference.
The code for the Impress server option is 10, and the minimum length for this
option is 4 octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 10 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.12 Option 11: Resource location server
This option specifies a list of RFC 887 resource location servers available to
the client. Servers should be listed in order of preference.
The code for this option is 11, and the minimum length for this option is 4
octets. The length must always be a multiple of 4.
Code Length
Address 1
Address 2 [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 11 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.13 Option 12: Host name
This option specifies the name of the client. The name may or may not
include the full local domain name. (See D.2.16, “Option 15: Domain name”
on page 462, for the preferred method of obtaining the domain name.)
RFC 1035 explains the restrictions concerning which characters can be used
in the host name.
The code for this option is 12, and its minimum length is 1.
Code Length
Host Name
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 12 | n | h1 | h2 | h3 | h4 | h5 | h6 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
Appendix D. DHCP options (RFC 2132)
461
D.2.14 Option 13: Boot file size
This option specifies the length, in 512-octet blocks, of the default boot image
file for the client. The file length is specified as an unsigned 16-bit integer.
The code for this option is 13, and its length is 2.
Code Length
File Size
+-----+-----+-----+-----+
| 13 | 2 | l1 | l2 |
+-----+-----+-----+-----+
D.2.15 Option 14: Merit dump file name
This option specifies the file name (with path) of a file where the client's core
image should be dumped in the event the client crashes. The path is
formatted as a string consisting of characters from the NVT ASCII character
set.
The code for this option is 14, and its minimum length is 1.
Code Length
Dump Path/File Name
+-----+-----+-----+-----+-----+-----+-----+
| 14 | n | n1 | n2 | n3 | n4 | [...]
+-----+-----+-----+-----+-----+-----+-----+
D.2.16 Option 15: Domain name
This option specifies the default domain name that the client should use when
resolving host names using the domain name system.
The code for this option is 15, and its minimum length is 1.
Code Length
Domain Name
+-----+-----+-----+-----+-----+-----+-----+
| 15 | n | d1 | d2 | d3 | d4 | [...]
+-----+-----+-----+-----+-----+-----+-----+
D.2.17 Option 16: Swap server
This specifies the IP address of the client's swap server.
The code for this option is 16, and its length is 4.
Code Length Swap Server Address
+-----+-----+-----+-----+-----+-----+
| 16 | n | a1 | a2 | a3 | a4 |
+-----+-----+-----+-----+-----+-----+
462
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
D.2.18 Option 17: Root path
This option specifies the path name that contains the client's root directory.
The path is formatted as a string consisting of characters from the NVT ASCII
character set.
The code for this option is 17, and its minimum length is 1.
Code Length
Root Disk Pathname
+-----+-----+-----+-----+-----+-----+-----+
| 17 | n | n1 | n2 | n3 | n4 | [...]
+-----+-----+-----+-----+-----+-----+-----+
D.2.19 Option 18: Extensions path
This option specifies the path and name of a file, retrievable through TFTP,
which contains information to be interpreted in the same way as the 64-octet
vendor extension field within the BOOTP response. The following exceptions
apply:
• The length of the file is not limited.
• All references to Tag 18 (instances of the BOOTP extensions path field)
within the file are ignored.
The code for this option is 18, and its minimum length is 1.
Code Length
Extensions Path/File Name
+-----+-----+-----+-----+-----+-----+-----+
| 18 | n | n1 | n2 | n3 | n4 | [...]
+-----+-----+-----+-----+-----+-----+-----+
D.2.20 Option 19: IP forwarding enable/disable
This option specifies whether the client should configure its IP layer for packet
forwarding. A value of 0 means disable IP forwarding, and a value of 1 means
enable IP forwarding.
The code for this option is 19, and its length is 1.
Code Length Value
+-----+-----+-----+
| 19 | 1 | n |
+-----+-----+-----+
D.2.21 Option 20: Non-local source routing enable/disable
This option specifies whether the client should configure its IP layer to allow
forwarding of datagrams with non-local source routes. A value of 0 means
Appendix D. DHCP options (RFC 2132)
463
disallow forwarding of such datagrams, and a value of 1 means allow
forwarding.
The code for this option is 20, and its length is 1.
Code Length Value
+-----+-----+-----+
| 20 | 1 | n |
+-----+-----+-----+
D.2.22 Option 21: Policy filter
This option specifies policy filters for non-local source routing. The filters
consist of a list of IP addresses and masks that specify destination/mask
pairs with which to filter incoming source routes.
Any source routed datagram whose next hop address does not match one of
the filters should be discarded by the client.
The code for this option is 21, and the minimum length of this option is 8. The
length must be a multiple of 8.
Code Length
Address 1
Mask 1
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 21 | n | a1 | a2 | a3 | a4 | m1 | m2 | m3 | m4 |
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
Address 2
Mask 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| a1 | a2 | a3 | a4 | m1 | m2 | m3 | m4 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.23 Option 22: Maximum datagram reassembly size
This option specifies the maximum size datagram that the client should be
prepared to reassemble. The size is specified as a 16-bit unsigned integer.
The minimum legal value is 576.
The code for this option is 22, and its length is 2.
Code Length Size
+-----+-----+-----+-----+
| 22 | 2 | s1 | s2 |
+-----+-----+-----+-----+
464
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
D.2.24 Option 23: Default IP time-to-live
This option specifies the default time-to-live that the client should use for
outgoing datagrams. The TTL is specified as an octet with a value between 1
and 255.
The code for this option is 23, and its length is 1.
Code Length TTL
+-----+-----+-----+
| 23 | 1 | ttl |
+-----+-----+-----+
D.2.25 Option 24: Path MTU aging timeout
This option specifies the timeout (in seconds) to use when aging path MTU
values discovered by the mechanism defined in RFC 1191. The timeout is
specified as a 32-bit unsigned integer.
The code for this option is 24, and its length is 4.
Code Length
Timeout
+-----+-----+-----+-----+-----+-----+
| 24 | 4 | t1 | t2 | t3 | t4 |
+-----+-----+-----+-----+-----+-----+
D.2.26 Option 25: Path MTU plateau table
This option specifies a table of MTU sizes to use when performing path MTU
discovery as defined in RFC 1191. The table is formatted as a list of 16-bit
unsigned integers, ordered from smallest to largest. The minimum MTU value
cannot be smaller than 68.
The code for this option is 25, and its minimum length is 2. The length must
be a multiple of 2.
Code Length Size 1
Size 2
+-----+-----+-----+-----+-----+-----+-----+
| 25 | n | s1 | s2 | s1 | s2 | [...]
+-----+-----+-----+-----+-----+-----+-----+
D.2.27 Option 26: Interface MTU
This option specifies the MTU to use for this interface. The MTU is specified
as a 16-bit unsigned integer. The minimum legal value for the MTU is 68.
Appendix D. DHCP options (RFC 2132)
465
The code for this option is 26, and its length is 2.
Code Length MTU
+-----+-----+-----+-----+
| 26 | 2 | m1 | m2 |
+-----+-----+-----+-----+
D.2.28 Option 27: All subnets are local
This option specifies whether or not the client may assume that all subnets of
the IP network to which the client is connected use the same MTU as the
subnet of that network to which the client is directly connected. A value of 1
indicates that all subnets share the same MTU. A value of 0 means that the
client should assume that some subnets of the directly connected network
may have smaller MTUs.
The code for this option is 27, and its length is 1.
Code Length Value
+-----+-----+-----+
| 27 | 1 | n |
+-----+-----+-----+
D.2.29 Option 28: Broadcast address
This option specifies the broadcast address in use on the client's subnet.
The code for this option is 28, and its length is 4.
Code Length Broadcast Address
+-----+-----+-----+-----+-----+-----+
| 28 | 4 | b1 | b2 | b3 | b4 |
+-----+-----+-----+-----+-----+-----+
D.2.30 Option 29: Perform mask discovery
This option specifies whether or not the client should perform subnet mask
discovery using ICMP. A value of 0 indicates that the client should not perform
mask discovery. A value of 1 means that the client should perform mask
discovery.
The code for this option is 29, and its length is 1.
Code Length Value
+-----+-----+-----+
| 29 | 1 | n |
+-----+-----+-----+
466
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
D.2.31 Option 30: Mask supplier
This option specifies whether or not the client should respond to subnet mask
requests using ICMP. A value of 0 indicates that the client should not
respond. A value of 1 means that the client should respond.
The code for this option is 30, and its length is 1.
Code Length Value
+-----+-----+-----+
| 30 | 1 | n |
+-----+-----+-----+
D.2.32 Option 31: Perform router discovery
This option specifies whether or not the client should solicit routers using the
router discovery mechanism defined in RFC 1256. A value of 0 indicates that
the client should not perform router discovery. A value of 1 means that the
client should perform router discovery.
The code for this option is 31, and its length is 1.
Code Length Value
+-----+-----+-----+
| 31 | 1 | n |
+-----+-----+-----+
D.2.33 Option 32: Router solicitation address
This option specifies the address to which the client should transmit router
solicitation requests.
The code for this option is 32, and its length is 4.
Code Length
Address
+-----+-----+-----+-----+-----+-----+
| 32 | 4 | a1 | a2 | a3 | a4 |
+-----+-----+-----+-----+-----+-----+
D.2.34 Option 33: Static route
This option specifies a list of static routes that the client should install in its
routing cache. If multiple routes to the same destination are specified, they
are listed in descending order of priority.
The routes consist of a list of IP address pairs. The first address is the
destination address, and the second address is the router for the destination.
The default route (0.0.0.0) is an illegal destination for a static route.
Appendix D. DHCP options (RFC 2132)
467
The code for this option is 33, and the minimum length for this option is 8. The
length must be a multiple of 8.
Code Length
Destination 1
Router 1
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 33 | n | d1 | d2 | d3 | d4 | r1 | r2 | r3 | r4 |
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
Destination 2
Router 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| d1 | d2 | d3 | d4 | r1 | r2 | r3 | r4 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.35 Option 34: Trailer encapsulation
This option specifies whether or not the client should negotiate the use of
trailers (RFC 893) when using the ARP protocol. A value of 0 indicates that
the client should not attempt to use trailers. A value of 1 means that the client
should attempt to use trailers.
The code for this option is 34, and its length is 1.
Code Length Value
+-----+-----+-----+
| 34 | 1 | n |
+-----+-----+-----+
D.2.36 Option 35: ARP cache timeout
This option specifies the timeout, in seconds, for ARP cache entries. The time
is specified as a 32-bit unsigned integer.
The code for this option is 35, and its length is 4.
Code Length
Time
+-----+-----+-----+-----+-----+-----+
| 35 | 4 | t1 | t2 | t3 | t4 |
+-----+-----+-----+-----+-----+-----+
D.2.37 Option 36: Ethernet encapsulation
This option specifies whether or not the client should use Ethernet Version 2
(RFC 894) or IEEE 802.3 (RFC 1042) encapsulation if the interface is
Ethernet. A value of 0 indicates that the client should use RFC 894
encapsulation. A value of 1 means that the client should use RFC 1042
encapsulation.
The code for this option is 36, and its length is 1.
468
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Code Length Value
+-----+-----+-----+
| 36 | 1 | n |
+-----+-----+-----+
D.2.38 Option 37: TCP default time-to-live
This option specifies the default time-to-live that the client should use when
sending TCP segments. The value is represented as an 8-bit unsigned
integer. The minimum value is 1.
The code for this option is 37, and its length is 1.
Code Length TTL
+-----+-----+-----+
| 37 | 1 | n |
+-----+-----+-----+
D.2.39 Option 38: TCP keepalive interval
This option specifies the interval, in seconds, that the client should wait
before sending a keepalive message on a TCP connection. The time is
specified as a 32-bit unsigned integer. A value of zero indicates that the client
should not generate keep-alive messages on connections unless specifically
requested by an application.
The code for this option is 38, and its length is 4.
Code Length
Time
+-----+-----+-----+-----+-----+-----+
| 38 | 4 | t1 | t2 | t3 | t4 |
+-----+-----+-----+-----+-----+-----+
D.2.40 Option 39: TCP keepalive garbage
This option specifies whether or not the client should send TCP keepalive
messages with an octet of garbage for compatibility with older
implementations. A value of 0 indicates that a garbage octet should not be
sent. A value of 1 indicates that a garbage octet should be sent.
The code for this option is 39, and its length is 1.
Code Length Value
+-----+-----+-----+
| 39 | 1 | n |
+-----+-----+-----+
Appendix D. DHCP options (RFC 2132)
469
D.2.41 Option 40: Network information service domain
This option specifies the name of the client's NIS domain. The domain is
formatted as a string consisting of characters from the NVT ASCII character
set.
The code for this option is 40, and its minimum length is 1.
Code Length
NIS Domain Name
+-----+-----+-----+-----+-----+-----+-----+
| 40 | n | n1 | n2 | n3 | n4 | [...]
+-----+-----+-----+-----+-----+-----+-----+
D.2.42 Option 41: Network information server
This option specifies a list of IP addresses indicating NIS servers available to
the client. Servers should be listed in order of preference.
The code for this option is 41, and its minimum length is 4. The length must
be a multiple of 4.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 41 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.43 Option 42: Network time protocol server
This option specifies a list of IP addresses indicating NTP servers available to
the client. Servers should be listed in order of preference.
The code for this option is 42, and its minimum length is 4. The length must
be a multiple of 4.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 42 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.44 Option 43: Vendor-specific information
This option is used by clients and servers to exchange vendor-specific
information. The information is an opaque object of n octets, presumably
interpreted by vendor-specific code on the clients and servers. The definition
of this information is vendor-specific.
The vendor is indicated in the vendor class identifier option. Servers not
equipped to interpret the vendor-specific information sent by a client must
ignore it, although it may be reported.
470
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Clients that do not receive desired vendor-specific information should make
an attempt to operate without it, although they may do so, and announce they
are doing so, with degraded functionality.
If a vendor potentially encodes more than one item of information in this
option, then the vendor should encode the option using encapsulated
vendor-specific options. The encapsulated vendor-specific options field
should be encoded as a sequence of code/length/value fields of identical
syntax to the DHCP options field with the following exceptions:
• There should not be a magic cookie field in the encapsulated
vendor-specific extensions field.
• Codes other than 0 or 255 may be redefined by the vendor within the
encapsulated vendor-specific extensions field, but should conform to the
tag/length/value syntax.
• Code 255 (End Option), if present, signifies the end of the encapsulated
vendor extensions, not the end of the vendor extensions field. If no code
255 is present, then the end of the enclosing vendor-specific information
field is taken as the end of the encapsulated vendor-specific extensions
field.
The code for this option is 43, and its minimum length is 1.
Code Length Vendor-Specific Information
+-----+-----+-----+-----+-----+
| 43 | n | i1 | i2 | [...]
+-----+-----+-----+-----+-----+
When encapsulated vendor-specific extensions are used, the information
bytes (i1 to in ) have the following format:
Code Length Data Item Code Length Data Item Code
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| T1 | n | d1 | d2 | ... | T2 | n | D1 | D2 | ... | ... |
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.45 Option 44: NetBIOS over TCP/IP name server option
The NetBIOS name server (NBNS) option specifies a list of RFC 1001/1002
NBNS name servers, such as Shadow IPserver, listed in order of preference.
Appendix D. DHCP options (RFC 2132)
471
The code for this option is 44, and the minimum length of the option is 4
octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 44 | n | a1 | a2 | a3 | a4 | b1 | b2 | b3 | b4 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.46 Option 45: NetBIOS over TCP/IP datagram distribution server
The NetBIOS datagram distribution server (NBDD) option specifies a list of
RFC 1001/1002 NBDD servers listed in order of preference.
The code for this option is 45, and the minimum length of the option is 4
octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 45 | n | a1 | a2 | a3 | a4 | b1 | b2 | b3 | b4 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.47 Option 46: NetBIOS over TCP/IP node type
The NetBIOS node type option allows NetBIOS over TCP/IP clients to be
configured as described in RFC 1001/1002. The value is specified as a single
octet that identifies the client type as follows:
Value (Hex)
----01
02
04
08
Node Type
--------B-node
P-node
M-node
H-node
The code for this option is 46, and the length of this option is always 1.
Code Length Node Type
+-----+-----+-----------+
| 46 | 1 | See Above |
+-----+-----+-----------+
D.2.48 Option 47: NetBIOS over TCP/IP scope
This option specifies the NetBIOS over TCP/IP scope parameter for the client
as described in RFC 1001/1002. (Certain restrictions on the characters that
can be used may apply.)
472
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The code for this option is 47, and the minimum length of this option is 1.
Code Length
NetBIOS Scope
+-----+-----+-----+-----+-----+-----+-----+
| 47 | n | s1 | s2 | s3 | s4 | [...]
+-----+-----+-----+-----+-----+-----+-----+
D.2.49 Option 48: X-Window system font server option
This option specifies a list of X-Window system font servers available to the
client. Servers should be listed in order of preference.
The code for this option is 48, and the minimum length of this option is 4
octets. The length must be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 48 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.50 Option 49: X-Window system display manager
This option specifies a list of IP addresses of systems that are running the X
Window System Display Manager and are available to the client. Addresses
should be listed in order of preference.
The code for this option is 49, and the minimum length of this option is 4. The
length must be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 49 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.51 Option 64: Network Information Service (plus) domain
This option specifies the name of the client's NIS+ domain. The domain is
formatted as a string consisting of characters from the NVT ASCII character
set.
The code for this option is 64, and its minimum length is 1.
Code Length
NIS+ Domain Name
+-----+-----+-----+-----+-----+-----+-----+
| 64 | n | n1 | n2 | n3 | n4 | [...]
+-----+-----+-----+-----+-----+-----+-----+
Appendix D. DHCP options (RFC 2132)
473
D.2.52 Option 65: Network Information Service (plus) server
This option specifies a list of IP addresses indicating NIS+ servers available
to the client. Servers should be listed in order of preference.
The code for this option is 65, and its minimum length is 4. The length must
be a multiple of 4.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 65 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.53 Option 68: Mobile IP home agent
This option specifies a list of IP addresses indicating mobile IP home agents
available to the client. Agents should be listed in order of preference.
The code for this option is 68. Its minimum length is 0, indicating no home
agents are available. The length must be a multiple of 4. The usual length will
be 4 octets, containing a single home agent's address.
Code Length
Address 1
+-----+-----+-----+-----+-----+-----+-----+
| 68 | n | a1 | a2 | a3 | a4 | [...]
+-----+-----+-----+-----+-----+-----+-----+
D.2.54 Option 69: Simple Mail Transport Protocol (SMTP) server
This option specifies a list of SMTP servers available to the client. Servers
should be listed in order of preference.
The code for the SMTP server option is 69, and the minimum length for this
option is 4 octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 69 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.55 Option 70: Post Office Protocol (POP3) server
This option specifies a list of POP3 mail servers available to the client.
Servers should be listed in order of preference.
474
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The code for the POP3 server option is 70, and the minimum length for this
option is 4 octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 70 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.56 Option 71: Network News Transport Protocol (NNTP) server
This option specifies a list of NNTP servers available to the client. Servers
should be listed in order of preference.
The code for the NNTP server option is 71, and the minimum length for this
option is 4 octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 71 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.57 Option 72: Default World Wide Web (WWW) server
This option specifies a list of WWW servers available to the client. Servers
should be listed in order of preference.
The code for the WWW server option is 72, and the minimum length for this
option is 4 octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 72 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.58 Option 73: Default Finger server
This option specifies a list of Finger servers available to the client. Servers
should be listed in order of preference.
The code for the Finger server option is 73, and the minimum length for this
option is 4 octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 73 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
Appendix D. DHCP options (RFC 2132)
475
D.2.59 Option 74: Default Internet Relay Chat (IRC) server
This option specifies a list of IRC servers available to the client. Servers
should be listed in order of preference.
The code for the IRC server option is 74, and the minimum length for this
option is 4 octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 74 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.60 Option 75: StreetTalk server
This option specifies a list of StreetTalk servers available to the client.
Servers should be listed in order of preference.
The code for the StreetTalk server option is 75 and the minimum length for
this option is 4 octets. The length must always be a multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 75 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.2.61 Option 76: StreetTalk Directory Assistance (STDA) server
This option specifies a list of StreetTalk Directory Assistance servers
available to the client. Servers should be listed in order of preference.
The code for the StreetTalk Directory Assistance server option is 76, and the
minimum length for this option is 4 octets. The length must always be a
multiple of four.
Code Length
Address 1
Address 2
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
| 76 | n | a1 | a2 | a3 | a4 | a1 | a2 | [...]
+-----+-----+-----+-----+-----+-----+-----+-----+-----+
D.3 DHCP (only) options
This section introduces all DHCP-only options that can be defined at the
DHCP server and delivered to requesting DHCP clients. Be aware that
although most DHCP servers can deliver all options illustrated here to
requesting DHCP clients, there are platform-specific dependencies that
determine whether or not requesting DHCP clients can actually make use of
all delivered information.
476
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
D.3.1 Option 50: Requested IP address
This option is used in a client request (DHCPDISCOVER) to allow the client
to request that a particular IP address be assigned.
The code for this option is 50, and its length is 4.
Code Length
Address
+-----+-----+-----+-----+-----+-----+
| 50 | 4 | a1 | a2 | a3 | a4 |
+-----+-----+-----+-----+-----+-----+
D.3.2 Option 51: IP address lease time
This option is used in a client request (DHCPDISCOVER or DHCPREQUEST)
to allow the client to request a lease time for the IP address. In a server reply
(DHCPOFFER), a DHCP server uses this option to specify the lease time it is
willing to offer. The time is expressed in seconds and is specified as a 32-bit
unsigned integer.
The code for this option is 51, and its length is 4.
Code Length
Lease Time
+-----+-----+-----+-----+-----+-----+
| 51 | 4 | t1 | t2 | t3 | t4 |
+-----+-----+-----+-----+-----+-----+
D.3.3 Option 52: Option overload
This option is used to indicate that the DHCP sname or file fields are being
overloaded by using them to carry DHCP options. A DHCP server inserts this
option if the returned parameters will exceed the usual space allotted for
options. If this option is present, the client interprets the specified additional
fields after it finishes interpretation of the standard option fields.
The code for this option is 52, and its length is 1. Legal values for this option
are:
Value
----1
2
3
Meaning
------the "file" field is used to hold options
the "sname" field is used to hold options
both fields are used to hold options
Code Length Value
+-----+-----+-----+
| 52 | 1 | n |
+-----+-----+-----+
Appendix D. DHCP options (RFC 2132)
477
D.3.4 Option 53: DHCP message type
This option is used to convey the type of the DHCP message.
The code for this option is 53, and its length is 1. Legal values for this option
are:
Value
----1
2
3
4
5
6
7
8
Message Type
-----------DHCPDISCOVER
DHCPOFFER
DHCPREQUEST
DHCPDECLINE
DHCPACK
DHCPNAK
DHCPRELEASE
DHCPINFORM
Code Length Type
+-----+-----+-----+
| 53 | 1 | n |
+-----+-----+-----+
D.3.5 Option 54: Server identifier
This option is used in DHCPOFFER and DHCPREQUEST messages and
may optionally be included in the DHCPACK and DHCPNAK messages.
DHCP servers include this option in the DHCPOFFER in order to allow the
client to distinguish between lease offers. DHCP clients use the contents of
the server identifier field as the destination address for any DHCP messages
unicast to the DHCP server. DHCP clients also indicate which of several
lease offers is being accepted by including this option in a DHCPREQUEST
message. The identifier is the IP address of the selected server.
The code for this option is 54, and its length is 4.
Code Length
Address
+-----+-----+-----+-----+-----+-----+
| 54 | 4 | a1 | a2 | a3 | a4 |
+-----+-----+-----+-----+-----+-----+
D.3.6 Option 55: Parameter request list
This option is used by a DHCP client to request values for specified
configuration parameters. The list of requested parameters is specified as (n)
octets, where each octet is a valid DHCP option code as defined in this
appendix.
478
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
The client may list the options in order of preference. The DHCP server is not
required to return the options in the requested order but must try to insert the
requested options in the order requested by the client.
The code for this option is 55, and its minimum length is 1.
Code Length Option Codes
+-----+-----+-----+-----+-----+
| 55 | n | c1 | c2 | [...]
+-----+-----+-----+-----+-----+
D.3.7 Option 56: Message
This option is used by a DHCP server to provide an error message to a DHCP
client in a DHCPNAK message in the event of a failure. A client may use this
option in a DHCPDECLINE message to indicate why the client declined the
offered parameters. The message consists of (n) octets of NVT ASCII text,
which the client may display on an available output device.
The code for this option is 56, and its minimum length is 1.
Code Length Text
+-----+-----+-----+-----+-----+
| 56 | n | c1 | c2 | [...]
+-----+-----+-----+-----+-----+
D.3.8 Option 57: Maximum DHCP message size
This option specifies the maximum length of a DHCP message that a system
is willing to accept. The length is specified as an unsigned 16-bit integer. A
client may use the maximum DHCP message size option in DHCPDISCOVER
or DHCPREQUEST messages, but should not use the option in
DHCPDECLINE messages.
The code for this option is 57, and its length is 2. The minimum legal value is
576 octets.
Code Length Length
+-----+-----+-----+-----+
| 57 | 2 | l1 | l2 |
+-----+-----+-----+-----+
D.3.9 Option 58: Renewal (T1) time value
This option specifies the time interval from address assignment until the client
transitions to the RENEWING state. The value is in seconds and is specified
as a 32-bit unsigned integer.
Appendix D. DHCP options (RFC 2132)
479
The code for this option is 58, and its length is 4.
Code Length T1 Interval
+-----+-----+-----+-----+-----+-----+
| 58 | 4 | t1 | t2 | t3 | t4 |
+-----+-----+-----+-----+-----+-----+
D.3.10 Option 59: Rebinding (T2) time value
This option specifies the time interval from address assignment until the client
transitions to the REBINDING state. The value is in seconds and is specified
as a 32-bit unsigned integer.
The code for this option is 59, and its length is 4.
Code Length T2 Interval
+-----+-----+-----+-----+-----+-----+
| 59 | 4 | t1 | t2 | t3 | t4 |
+-----+-----+-----+-----+-----+-----+
D.3.11 Option 60: Vendor class identifier
This option is used by DHCP clients to optionally identify the vendor type and
configuration of a DHCP client. The information is a string of (n) octets,
interpreted by servers. Vendors may choose to define specific vendor class
identifiers to convey particular configuration or other identification information
about a client. For example, the identifier may include the client's hardware
configuration. Servers not equipped to interpret the class-specific information
sent by a client must ignore it, although it may be reported. Servers that
respond should only use option 43 to return the vendor-specific information to
the client.
The code for this option is 60, and its minimum length is 1.
Code Length Vendor Class Identifier
+-----+-----+-----+-----+-----+
| 60 | n | i1 | i2 | [...]
+-----+-----+-----+-----+-----+
D.3.12 Option 61: Client identifier
This option is used by DHCP clients to specify their own unique identifiers.
DHCP servers use this value to index their database of address bindings.
This value is expected to be unique for all clients in an administrative domain.
Identifiers should be treated as opaque objects by DHCP servers. The client
identifier may consist of type-value pairs similar to htype/chaddr fields. For
instance, it may consist of a hardware type and hardware address. In this
480
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
case, the type field should be a defined ARP hardware type. A hardware type
of 0 (zero) should be used when the value field contains an identifier other
than a hardware address (for instance, a fully qualified domain name).
For correct identification of clients, each client's identifier must be unique
among the identifiers used on the subnet to which the client is attached.
Vendors and system administrators are responsible for choosing identifiers
that meet this requirement for uniqueness.
The code for this option is 61, and its minimum length is 2.
Code Length Type Client Identifier
+-----+-----+-----+-----+-----+-----+
| 61 | n | t1 | i1 | i2 | [...]
+-----+-----+-----+-----+-----+-----+
D.3.13 Option 66: TFTP server name
This option is used to identify a TFTP server when the sname field in the
DHCP header has been used for DHCP options.
The code for this option is 66, and its minimum length is 1.
Code Length
TFTP server
+-----+-----+-----+-----+-----+-----+
| 66 | n | c1 | c2 | c3 | [...]
+-----+-----+-----+-----+-----+-----+
D.3.14 Option 67: Boot file name
This option is used to identify a boot file when the file field in the DHCP
header has been used for DHCP options.
The code for this option is 67, and its minimum length is 1.
Code Length Boot File Name
+-----+-----+-----+-----+-----+-----+
| 67 | n | c1 | c2 | c3 | [...]
+-----+-----+-----+-----+-----+-----+
D.4 Unofficial DHCP options
Several additional options, while not yet part of the official RFC 2132
standard, have achieved some degree of popularity and are in common use.
If you are interested in the latest inventory of unofficially assigned DHCP
options, the list is available on the Internet at:
ftp://ftp.isi.edu/in-notes/iana/assignments
Appendix D. DHCP options (RFC 2132)
481
Option
62
63
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
126
127
482
Description
NetWare/IP Domain Name
NetWare/IP Suboptions
User Class
Directory Agent Information
Service Location Agent Scope
Naming Authority
Client Fully Qualified Domain Name
Agent Circuit ID
Agent Remote ID
Agent Subnet Mask
Novell Directory Services Servers
Novell Directory Services Tree Name
Novell Directory Services Context
IEEE 1003.1 POSIX Timezone
Fully Qualified Domain Name
Authentication
Banyan Vines TCP/IP Server
Server Selection
Client System Architecture
Client Network Device Interface
Lightweight Directory Access Protocol (LDAP)
IPv6 Transitions
UUID/GUID-based Client Identifier
Printer Name
Multicast Scope
Start Time
Multicast Time-to-Live
Multicast Block Size
Client Port
Cookie
Multicast Scope List
Swap Path
Autonomous System Number (ASN)
IPX Compatibility
Served IP Range
NetInfo Parent Server Address
NetInfo Parent Server Tag
Universal Resource Locator (URL)
DHCP Failover Protocol
Extension
Extension
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
By convention, OS/2 Warp (and likely other IBM products) may use additional
DHCP options in the 200 to 208 range. You may wish to use these same
option numbers for the same purposes, even on non-IBM platforms.
200
201
202
203
204
205
206
207
208
Default LPR Printer
Default Gopher Server
Default WWW Home Page
Default WWW Proxy Gateway
Default WWW News Server
Default SOCKS Server
NFS Mount Points
Default X Font Servers
X Display Manager Servers
D.5 Options supported by popular operating systems
This section provides information about strengths and weaknesses some
DHCP servers and DHCP clients encounter.
D.5.1 Servers
OS/2 Warp Server, AIX, Shadow IPserver, and Windows NT Server can serve
any and all DHCP options to clients. However, it appears that it is not possible
to alter Option 1 in at least some releases of Windows NT Server in order to
assign a subnet mask to a client which is not identical to the server’s own
subnet mask.
OS/2 Warp Server and AIX tend to provide the most flexibility in defining
address pools.
D.5.2 Clients
OS/2 Warp and WorkSpace On-Demand support all DHCP options, and
DHCP option handlers are provided for most of the standard DHCP options.
Others can be interpreted and acted upon by the client using custom option
handlers. (See 7.2, “Roaming users” on page 322 for some examples.)
Windows 95 and Windows NT support a more limited set of options, notably
1, 3, 6, 44, 46, and 47. Third-party products may be required to intercept and
handle less commonly used DHCP options.
Appendix D. DHCP options (RFC 2132)
483
484
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Appendix E. Special notices
This publication is intended to help TCP/IP network operators establish a
solid foundation for future growth without undue expense. Enterprise network
staff will appreciate the sections exploring high-end technologies, including
Network TeleSystems’ Shadow IPserver, IBM Communications Server, and
UNIX platforms. All network managers should find the authors’ advice on
security well worth reading.
References in this publication to IBM products, programs or services do not
imply that IBM intends to make these available in all countries in which IBM
operates. Any reference to an IBM product, program, or service is not
intended to state or imply that only IBM's product, program, or service may be
used. Any functionally equivalent program that does not infringe any of IBM's
intellectual property rights may be used instead of the IBM product, program
or service.
Information in this book was developed in conjunction with use of the
equipment specified, and is limited in application to those specific hardware
and software products and levels.
IBM may have patents or pending patent applications covering subject matter
in this document. The furnishing of this document does not give you any
license to these patents. You can send license inquiries, in writing, to the IBM
Director of Licensing, IBM Corporation, North Castle Drive, Armonk, NY
10504-1785.
Licensees of this program who wish to have information about it for the
purpose of enabling: (i) the exchange of information between independently
created programs and other programs (including this one) and (ii) the mutual
use of the information which has been exchanged, should contact IBM
Corporation, Dept. 600A, Mail Drop 1329, Somers, NY 10589 USA.
Such information may be available, subject to appropriate terms and
conditions, including in some cases, payment of a fee.
The information contained in this document has not been submitted to any
formal IBM test and is distributed AS IS. The information about non-IBM
("vendor") products in this manual has been supplied by the vendor and IBM
assumes no responsibility for its accuracy or completeness. The use of this
information or the implementation of any of these techniques is a customer
responsibility and depends on the customer's ability to evaluate and integrate
them into the customer's operational environment. While each item may have
been reviewed by IBM for accuracy in a specific situation, there is no
© Copyright IBM Corp. 1998, 2000
485
guarantee that the same or similar results will be obtained elsewhere.
Customers attempting to adapt these techniques to their own environments
do so at their own risk.
Any pointers in this publication to external Web sites are provided for
convenience only and do not in any manner serve as an endorsement of
these Web sites.
Any performance data contained in this document was determined in a
controlled environment, and therefore, the results that may be obtained in
other operating environments may vary significantly. Users of this document
should verify the applicable data for their specific environment.
This document contains examples of data and reports used in daily business
operations. To illustrate them as completely as possible, the examples
contain the names of individuals, companies, brands, and products. All of
these names are fictitious and any similarity to the names and addresses
used by an actual business enterprise is entirely coincidental.
Reference to PTF numbers that have not been released through the normal
distribution process does not imply general availability. The purpose of
including these reference numbers is to alert IBM customers to specific
information relative to the implementation of the PTF when it becomes
available to each customer according to the normal IBM PTF distribution
process.
The following terms are trademarks of the International Business Machines
Corporation in the United States and/or other countries:
AS/400
CT
eNetwork
LAN Distance
Netfinity Manager
Nway
Presentation Manager
SecureWay
System/390
XT
AT
DB2
IBM
Netfinity
Network Station
OS/2
RS/6000
SP
ThinkPad
400
The following terms are trademarks of other companies:
Tivoli, Manage. Anything. Anywhere.,The Power To Manage., Anything.
Anywhere.,TME, NetView, Cross-Site, Tivoli Ready, Tivoli Certified, Planet
Tivoli, and Tivoli Enterprise are trademarks or registered trademarks of Tivoli
Systems Inc., an IBM company, in the United States, other countries, or both.
486
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
In Denmark, Tivoli is a trademark licensed from Kjøbenhavns Sommer - Tivoli
A/S.
C-bus is a trademark of Corollary, Inc. in the United States and/or other
countries.
Java and all Java-based trademarks and logos are trademarks or registered
trademarks of Sun Microsystems, Inc. in the United States and/or other
countries.
Microsoft, Windows, Windows NT, and the Windows logo are trademarks of
Microsoft Corporation in the United States and/or other countries.
PC Direct is a trademark of Ziff Communications Company in the United
States and/or other countries and is used by IBM Corporation under license.
ActionMedia, LANDesk, MMX, Pentium and ProShare are trademarks of Intel
Corporation in the United States and/or other countries.
UNIX is a registered trademark in the United States and other countries
licensed exclusively through The Open Group.
SET and the SET logo are trademarks owned by SET Secure Electronic
Transaction LLC.
Other company, product, and service names may be trademarks or service
marks of others.
Appendix E. Special notices
487
488
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Appendix F. Related publications
The publications listed in this section are considered particularly suitable for a
more detailed discussion of the topics covered in this redbook.
F.1 IBM Redbooks publications
For information on ordering these publications see “How to get IBM
Redbooks” on page 495.
• OS/2 Warp Server, Windows NT, and NetWare: A Network Operating
System Study , SG24-4786
• AIX and Windows NT: Solutions for Interoperability, SG24-5102
• WorkSpace On-Demand Handbook Release 2.0, SG24-5117
• WorkSpace On-Demand 2.0 Feature for Windows Clients, SG24-5396
• AS/400 TCP/IP Autoconfiguration: DNS and DHCP Support, SG24-5147
• A Comprehensive Guide to Virtual Private Networks, Vol.1: IBM Firewall,
Server and Client Solutions, SG24-5201
• TCP/IP Tutorial and Technical Overview, GG24-3376
• Using IBM Firewall 3.1 for AIX, SG24-2577
• Check Point FireWall-1 on AIX A Cookbook for Stand-Alone and High
Availability, SG24-5492
• A Comprehensive Guide to Virtual Private Networks, Volume II: IBM
Nways Router Solutions, SG24-5234
• A Comprehensive Guide to Virtual Private Networks, Volume III: IBM
Cross-Platform and Key Management Solutions, SG24-5309
• All About Tivoli Management Agents, SG24-5134
• TME 10 Internals and Problem Determination, SG24-2034
• Learning Practical TCP/IP for AIX v3.2/v4.1 Users: Hints and Tips for
Debugging and Tuning, SG24-4381
• HACMP Enhanced Scalability Handbook, SC24-5328
• IBM TCP/IP Performance Tuning Guide, SC31-7188
The following publications are available in softcopy format at
http://www.redbooks.ibm.com/ and may be available in hardcopy format at
www.fatbrain.com:
© Copyright IBM Corp. 1998, 2000
489
• Network Clients for OS/2 Warp Server: OS/2 Warp4, DOS/Windows,
Windows 95/NT, and Apple Macintosh, SG24-2009
• The OS/2 Warp 4 CID Software Distribution Guide, SG24-2010
• The OS/2 Warp 4 CID Rapid Deployment Tools Migration and Installation
Scenarios, SG24-2012
• TCP/IP Implementation in an OS/2 Warp Environment, SG24-4730
F.2 IBM Redbooks collections
Redbooks are also available on the following CD-ROMs. Click the CD-ROMs
button at http://www.redbooks.ibm.com/ for information about all the CD-ROMs
offered, updates and formats.
CD-ROM Title
System/390 Redbooks Collection
Networking and Systems Management Redbooks Collection
Transaction Processing and Data Management Redbooks Collection
Lotus Redbooks Collection
Tivoli Redbooks Collection
AS/400 Redbooks Collection
Netfinity Hardware and Software Redbooks Collection
RS/6000 Redbooks Collection (BkMgr)
RS/6000 Redbooks Collection (PDF Format)
Application Development Redbooks Collection
IBM Enterprise Storage and Systems Management Solutions
Collection Kit
Number
SK2T-2177
SK2T-6022
SK2T-8038
SK2T-8039
SK2T-8044
SK2T-2849
SK2T-8046
SK2T-8040
SK2T-8043
SK2T-8037
SK3T-3694
F.3 Other resources
These publications are also relevant as further information sources:
• Internetworking with TCP/IP, Volume I, Principles, Protocols and
Architecture, third edition, Prentice-Hall, Inc., ISBN 0-13-216987-8
• TCP/IP Tutorial and Technical Overview, Prentice-Hall, Inc., ISBN
0-13-460858-5
• IPng and the TCP/IP Protocols, John Wiley & Sons, Inc., ISBN
0-471-13088-5
• Communications for Cooperating Systems - OSI, SNA and TCP/IP,
Addison-Wesley Publishing Company, Inc., ISBN 0-201-50775-7
• DNS and BIND, O’Reilly & Associates, ISBN 1-56592-236-0
490
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
• Sendmail (Nutshell Handbook), O’Reilly & Associates, ISBN
1-56592-222-0
• Internetworking with Microsoft TCP/IP in Microsoft Windows NT 4.0,
Microsoft 0871B
• TCP/IP Architecture, Protocols and Implementation with IP v6 and IP
Security, McGraw-Hill, 0-07-021389-5
• Inside TCP/IP Third Edition, New Riders, 1-56205-74-6
• Mastering TCP/IP for NT Server, Sybex, 0-7821-2123-3
• DNS and BIND, 2nd Edition; O’Reilly & Associates,ISBN: 1-56592-236-0
• RFC 1034 Domain Names - Concepts and Facilities
• RFC 1035 Domain Names - Implementation and Specification
F.4 Referenced Web sites
• IBM Software Choice:
http://www.ibm.com/software/os/warp/swchoice/
http://www.software.ibm.com/swchoice
• IBM Software Support:
http://ps.software.ibm.com
• IBM PC-DOS 2000:
http://www.software.ibm.com/os/dos
• IBM Global Services:
http://www.ibm.com/services/
• IBM SecureWay Communications Server:
http://www.software.ibm.com/enetwork/commserver
• IBM OS/2 Warp Device Driver Pak On-Line:
http://service.software.ibm.com/os2ddpak/index.htm
• IBM Hursley Labs (Java fixes):
ftp://ftp.hursley.ibm.com
ftp://ftp.hursley.ibm.com/pub/java/fixes/os2/11/114
• http://www.redbooks.ibm.com
• http://www.thursby.com
• http://www.redhat.com/corp/support/manuals/index.html
• http://www.redhat.com/community/user_groups.html
Appendix F. Related publications
491
• Software for implementing Windows with IBM Dynamic IP Client:
http://techsupport.service.ibm.com/asd-bin/doc/en_us/win95cl/f-feat.htm
• IBM OS2PopS Home Page:
http://www.raleigh.ibm.com/misc/os2pops
• IBM Web Traffic Express (evaluation version):
http://www.software.ibm.com/webserver/wte/index.htm
• Information on configuring AIX 4.3.3 as a DNS server:
http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/aixbman/commadmn/tcp
_nameres.htm#BCAF2DF137dada
• Domino Go Webserver:
http://www.software.ibm.com/webservers/dgw/prodlist.htm
• Information on named.conf:
http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/files/aixfiles/named
.conf.htm
• CIFS Internet draft:
http://search.ietf.org/internet-drafts/draft-leach-cifs-v1-spec-02.txt
http://www.cifs.com/spec.html
• Neighborhood Browser Enabler for OS/2 Warp Server Version 4:
http://service.software.ibm.com/asd-bin/doc/index.htm
• F/X Communications (InJoy, InJoy Connect, and Tunnel/2):
http://www.fx.dk
• Network TeleSystems (Shadow IPserver):
http://www.nts.com/
• Information on NETAnywhere:
http://vrcomm.com/html/vcpos2.html
• Computer Emergency Respose Team (CERT):
http://www.cert.org
http://www.cert.org/advisories
• Interaction between DHCP and DNS:
http://www.ietf.org/internet-drafts/draft-ietf-dhc-dhcp-dns-08.txt
• Vinca’s StandbyServer for OS/2 Warp:
http://www.vinca.com/products/os2/os2_sbs.html
• HP JetDirect Print Server Software Installation Guide:
492
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
http://www.hp.com/cposupport/manindex/jd600t_en_man.html
• QIP demo software:
http://www.lucent.com/ins/products/qip.html
• "Hobbes" OS/2 Warp Program Library:
ftp://ftp-os2.nmsu.edu/pub/os2/util/archiver
ftp://ftp-os2.nmsu.edu/pub/os2/apps/internet/ftp/client
• IP08267 FixPak for Warp Server:
ftp://ftp.software.ibm.com/ps/products/lan/fixes/lsv5.0/english-us
• IBM DB2 Universal Database Information and Support:
http://www.software.ibm.com/data
http://www.software.ibm.com/data/db2/db2tech
http://www-4.ibm.com/software/data/db2db2tech/version61.html
• DHCP options:
ftp://ds.internic.net/rfc/rfc/rfc2142.txt
• Unofficial DHCP options:
ftp://ftp.isi.edu/in-notes/iana/assignments
• Institute of Electrical and Electronics Engineers:
http://www.ieee.org
• Internet Engineering Task Force:
http://www.ietf.org
F.5 Useful Web and FTP sites
• There are more than 2,500 RFCs today. For those readers who want to
keep up-to-date with the latest advances in TCP/IP, the ever-increasing
number of RFCs and Internet Drafts (IDs), published by the non-profit
Internet Engineering Task Force, are the best sources. RFCs can be
viewed on the Internet at:
http://www.rfc-editor.org/
• The DHCP Frequently Asked Questions List:
http://web.syr.edu/~jmwobus/comfaqs/dhcp.faq.html
• IBM REXX Information (including versions for Windows 95 and NT):
http://rexx.hursley.ibm.com/rexx/rexxibm.htm
• IBM Network Computing Software Updates:
http://ps.boulder.ibm.com/pbin-usa-ps/getobj.pl?/pdocs-usa/softupd.html
Appendix F. Related publications
493
• Internet Engineering Task Force RFCs:
http://www.rfc-editor.org/
• IBM OS/2 Warp TCP/IP V4.1 Stack Updates Page:
http://ps.software.ibm.com
• "Hobbes" OS/2 Warp Program Library:
ftp://ftp-os2.nmsu.edu/pub/os2
• Internet Software Consortium:
http://www.isc.com/
• Info-ZIP’s Home Page:
http://www.cdrom.com/pub/infozip
• Checkpoint Software:
http://www.checkpoint.com/
• The Fourth Crusade:
http://www.fourthcrusade.com
• Association for Computing Machinery:
http://www.acm.org
• LanOptics Corporation (provider of firewalls and more; OS/2 Warp version
available upon request):
http://www.lanoptics.com
• Opera Software (Web browser):
http://www.operasoft.com
• To reduce network traffic for everyone else, IBM employees should use the
following internal Web addresses.
IBM Network Computing Software Updates:
http://os2service.austin.ibm.com
494
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
How to get IBM Redbooks
This section explains how both customers and IBM employees can find out about IBM Redbooks,
redpieces, and CD-ROMs. A form for ordering books and CD-ROMs by fax or e-mail is also provided.
• Redbooks Web Site http://www.redbooks.ibm.com/
Search for, view, download, or order hardcopy/CD-ROM Redbooks from the Redbooks Web site.
Also read redpieces and download additional materials (code samples or diskette/CD-ROM images)
from this Redbooks site.
Redpieces are Redbooks in progress; not all Redbooks become redpieces and sometimes just a few
chapters will be published this way. The intent is to get the information out much quicker than the
formal publishing process allows.
• E-mail Orders
Send orders by e-mail including information from the IBM Redbooks fax order form to:
In United States
Outside North America
e-mail address
[email protected]
Contact information is in the “How to Order” section at this site:
http://www.elink.ibmlink.ibm.com/pbl/pbl
• Telephone Orders
United States (toll free)
Canada (toll free)
Outside North America
1-800-879-2755
1-800-IBM-4YOU
Country coordinator phone number is in the “How to Order”
section at this site:
http://www.elink.ibmlink.ibm.com/pbl/pbl
• Fax Orders
United States (toll free)
Canada
Outside North America
1-800-445-9269
1-403-267-4455
Fax phone number is in the “How to Order” section at this site:
http://www.elink.ibmlink.ibm.com/pbl/pbl
This information was current at the time of publication, but is continually subject to change. The latest
information may be found at the Redbooks Web site.
IBM Intranet for Employees
IBM employees may register for information on workshops, residencies, and Redbooks by accessing
the IBM Intranet Web site at http://w3.itso.ibm.com/ and clicking the ITSO Mailing List button.
Look in the Materials repository for workshops, presentations, papers, and Web pages developed
and written by the ITSO technical professionals; click the Additional Materials button. Employees may
access MyNews at http://w3.ibm.com/ for redbook, residency, and workshop announcements.
© Copyright IBM Corp. 1998, 2000
495
IBM Redbooks fax order form
Please send me the following:
Title
Order Number
First name
Quantity
Last name
Company
Address
City
Postal code
Country
Telephone number
Telefax number
VAT number
Card issued to
Signature
Invoice to customer number
Credit card number
Credit card expiration date
We accept American Express, Diners, Eurocard, Master Card, and Visa. Payment by credit card not
available in all countries. Signature mandatory for credit card payment.
496
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Glossary and abbreviations
ACM . Association for Computing Machinery.
Professional organization of computing
academics and other experts. See
http://www.acm.org on the Internet.
ADSL . Asynchronous Digital Subscriber Line.
Method of providing high-speed data transfer
over standard copper telephone wiring. Typically
offers higher performance than ISDN.
AFS. Andrew File System. One of three popular
network file systems, originating with UNIX, to
provide file and disk sharing across a TCP/IP
network. See also DFS and NFS.
AIX . IBM UNIX-compatible operating system for
the RS/6000 line of workstations and servers.
ANSI. American National Standards Institute.
Group which promulgates several different
technical standards, including the popular ANSI
terminal emulation standard.
APAR . A documented problem in an IBM
product, normally one for which a patch or other
fix is available.
AppleTalk . Network protocol used primarily to
connect computers and devices produced by
Apple Computer. AppleTalk is based on the
ISO/OSI Reference Model and incorporates the
SPX protocol.
ARP. Address Resolution Protocol. A protocol
used between routers and nodes to determine
the MAC or OSI physical layer address when the
Network layer (IP) address is known.
AS/400. Application System/400. An IBM line of
midrange computing systems, popular for
line-of-business applications.
ASCII . American Standard Code for Information
Interchange. A standard correspondence
between alphanumeric and control characters,
and (normally) 7-bit binary values (0 to 127). For
example, code 65 is defined as the uppercase
letter A in the ASCII table. Facilitates
communication between different computer
systems. Other common character set
© Copyright IBM Corp. 1998, 2000
definitions include Unicode (16-bit) and
EBCIDIC (8-bit).
ATM . Asynchronous Transfer Mode. It is a
high-speed (155 - 622 Mbps) communications
transport facility capable of carrying voice, data,
and video signaling.
BCP. Best Common Practice. An article by
which the IETF community can define and ratify
the Internet community's best current thinking
on a statement of principle or on what is
believed to be the best way to perform some
operations or IETF process function.
Binary. Having two components or possible
states. Usually represented by a code of zeros
and ones.
Boot Manager. Included with IBM OS/2 Warp
products and PowerQuest’s Partition Magic
software. Presents a menu on startup which
allows a PC user to choose which operating
system (OS/2 Warp, Windows 95, DOS,
Windows NT, and so on) to run among those
installed.
BootP. Bootstrap Protocol. An early method of
providing IP address information from a central
server based on a client’s network adapter
address. See also DHCP.
Bridge. A network relay that reads, buffers, and
forwards datagrams from one part of a network
to another. As a result, the two parts of a
network act as one. Bridges operate at the
Datalink layer of the OSI Model, or more
precisely, at the Media Access Control (MAC)
sublayer. See also Filter.
Broadcast. A transmission of a message (such
as a packet or frame) sent to all nodes on a
network rather than to a specific station.
Brouter. A single device which acts as both a
bridge and a router.
BSD. Berkeley Standard Distribution, a version
of UNIX.
497
Byte . Short for binary digit eight. A unit of
information consisting of usually eight bits. One
byte is sometimes referred as one octet.
Cable Modem. A device used to allow a
computer to communicate over cable television
coaxial wiring, usually to provide high-speed
Internet service.
CGI . Common Gateway Interface, a standard
mechanism used to pass data collected by a Web
server to a server-based application and receive
back any results, often used for "intelligent" Web
pages, such as order entry forms and credit card
processing.
Chooser. The application used on the Macintosh
to access AppleShare network services.
CIFS . Common Internet File System protocol.
CPSR . Computer Professionals for Social
Responsibility. See http://www.cpsr.org on the
Internet.
CSD. Corrective Service Diskette(s). Collection of
program fixes issued periodically for IBM
software products. Also known as Service Packs.
Daemon . A task that runs in the background,
usually to provide some kind of network service
to clients. For example, FTPD stands for FTP
Daemon.
Datagram. A packet of information transmitted
across a computer network consisting of a
header and one or more bytes of actual data
requiring no response or acknowledgment.
DDNS. Dynamic Domain Name System. See also
DNS.
DCE . Distributed Computing Environment. The
Open Software Foundation’s TCP/IP-based
protocols for security, DFS, and distributed
applications (RPC).
DFS. Distributed File System. One of three
popular network file systems, originating with
UNIX, to provide file and disk sharing across a
TCP/IP network. Part of the Open Software
Foundation’s DCE. See also AFS and NFS.
DHCP. The Dynamic Host Configuration Protocol
server provides IP addresses and IP
498
configuration information to IP addressrequesting workstations.
DLS. DOS LAN Services. IBM product for PCs
running DOS and/or Windows 3.x. Provides
network connection capabilities for file and printer
sharing. Also provides network protocol support,
including DHCP and basic TCP/IP services.
Included with Warp Server.
Domain . On the Internet, a domain is a part of
the naming hierarchy. The domain name is a
sequence of names (separated by periods) that
identify host sites. The leftmost part of a domain
name is the most specific, and the part on the
right is the most general.
DNS. The Domain Name System (DNS) is a
hierarchical, distributed method of organizing
systems and network names on the Internet.
DNS administratively groups hosts (systems) into
a hierarchy of authority that allows addressing
and other information to be widely distributed and
maintained. A big advantage of DNS is that using
it eliminates dependence on a
centrally-maintained file that maps host names to
IP addresses.
DOS. Disk Operating System. The software
programs that control the operation of the
computer and the movement of information
throughout the computer system. Although
operating systems known as DOS have been
produced for other systems, most people know
DOS as one of the popular operating systems for
the PC. First introduced in 1981 by Microsoft and
IBM, the newest version is called IBM PC-DOS
2000. See
http://www.software.ibm.com/os/dos for more
information.
Duplex. Pertaining to simultaneous two-way
independent data transmission in both directions
(as in full duplex).
E-mail. Electronic mail (e-mail) is the most
popular Internet application, and the driving force
behind the Internet’s rapid growth.
Ethernet. LAN datalink protocol developed by a
consortium of vendors, later standardized as
IEEE 802.3 with a few modifications. (For many
applications, users have not adopted all the IEEE
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
802.3 differences.) Ethernet 802.3 can now be
run on two types of coaxial cable as well as on
multi-mode fiber and unshielded twisted-pair
cables. The raw rate of data transmission is
typically 10 megabits/second.
Extranet . Using Internet technologies (such as
Web, Java, and TCP/IP) to link an internal
company network with networks operated by
vendors, suppliers, company partners, and
others, usually with at least some security. See
also Internet and intranet.
Fast Ethernet. A nickname for the 100 Mbps
version of IEEE 802.3.
FDDI. Fiber Distributed Data Interface (100 Mbps
fiber optic LAN).
File. A sequence of bytes stored on a secondary
storage medium, such as floppy disk or hard disk.
Filter. A network device or agent (program)
designed to separate data, signals, or material in
accordance with specified criteria.
Firewall. Generic term for a sophisticated filter
designed to connect a private network (intranet)
to a public network (Internet) in a secure way.
See also SOCKS and Proxy.
FTP. The File Transfer Protocol is part of the
TCP/IP protocol suite that is used to transfer files
between any two computers, provided they
support FTP. The two computers do not have to
be running the same operating system.
FQDN . Fully Qualified Domain Name. Includes
hostname plus domain name.
Gigabit Ethernet. Refers to Ethernet
technologies promising 1000 Mbps raw
transmission speed.
GUI . Graphical User Interface. A GUI uses
graphic representations of commands and/or a
menu format to display commands that the user
may execute with a mouse or similar device.
Host. Any computer attached to a TCP/IP
network that is remotely accessible through an
address.
HPFS . High Performance File System, part of
OS/2 Warp. The 32-bit version of this file system
is called HPFS386.
HTML. The HyperText Markup Language is the
standard language that the Web uses for creating
and recognizing hypermedia documents. Web
documents are most often written in HTML and
normally have an .html or .htm extension.
HTTP. HyperText Transport Protocol, the method
used by Web browsers to retrieve Web pages
from a Web server across a TCP/IP network.
Hub. A central point, or terminus, for network
connections or wiring.
Hyperlink . Words, phrases, images, or
characters highlighted in bold or underlined
indicate connections in a given document to
information within another document.
Hypermedia. Hypermedia is the name for richly
formatted documents containing a variety of
information types, such as textual, image, video,
and audio. These information types are easily
found through hyperlinks.
Hypertext. Hypertext allows users to move from
one site or place in a document to another.
Hypertext links in World Wide Web documents
link the user from terms in one document to the
site referenced in the original document.
IBM . International Business Machines.
ICMP. The Internet Control Message Protocol is
used for error reporting and recovery, and is a
required component of any IP implementation.
Described by RFC 792.
Icon. A graphical picture used to represent an
application, folder, file, disk drive, or printer.
IDE . Integrated Drive Electronics. This is a
standard interface for a hard disk drive or
CD-ROM drive. An IDE bus can be identified by
its 40-pin connector, as opposed to the 50-pin
connector of a SCSI bus.
IEEE . Institute of Electrical and Electronic
Engineers. A professional ANSI-accredited body
of scientists and engineers based in the United
States. See http://www.ieee.org on the
Internet.
499
IEEE 802 . The set of IEEE standards for the
definition of LAN protocols.
(installed network adapters). The IP address is
used to identify packet source and destination
host.
IEEE 802.2 . An IEEE standard describing the
parts of all IEEE LAN datalink protocols that are
common.
IPv6 . Internet Protocol Version 6, described by
RFCs 1883 through 1887.
IEEE 802.3 . An IEEE standard for LANs (an
improved version of Ethernet). See also Ethernet.
IPX . A network protocol developed by Novell for
use with its NetWare products.
IEEE 802.5 . An IEEE standard for token-ring
LANs. There are three types: 4 Mbps, 16 Mbps,
and 100 Mbps. See also Token-Ring.
IRC. Internet Relay Chat. A mechanism for
providing live messages across a TCP/IP
network, to allow individuals to converse in
groups.
IETF. Internet Engineering Task Force, a group
responsible for defining protocols in the TCP/IP
family for use on the Internet. See
http://www.ietf.org on the Internet.
IFS. Installable File System, a program used to
manage file storage on disks and other storage
devices. The program can be added or removed
from an operating system to change its file
storage behavior. For example, installable file
systems on OS/2 Warp include HPFS, HPFS386,
NFS, CD-ROM, and others.
IGMP. Internet Group Management Protocol,
maps IP addresses to hardware addresses on a
network. Also allows routers to check hosts to
see if they are interested in participating in
multicasts.
IMAP. Internet Message Access Protocol, used
primarily for TCP/IP-based e-mail applications.
Internet. The worldwide public network linking
computers using TCP/IP, Java, and other
standard networking technologies. Began in the
U.S. as a result of collaboration between the
Department of Defense, research institutions,
and universities. See also extranet and intranet.
Intranet. A private network established entirely
within a company or other organization based on
Internet technologies. See also extranet and
Internet.
IP. Internet Protocol, the basic protocol of TCP/IP
and of the Internet. IP is the OSI layer 3 routed
protocol used to transmit packetized information
on a TCP/IP network.
IP Address. Each host in the network is assigned
a unique IP address for each network connection
500
ISDN . Integrated Services Digital Network. A
digital telephone service used for high speed
Internet access and voice services. Delivered
over standard copper telephone wiring. See also
ADSL.
ISP. Internet Service Providers are companies
that provide an Internet connection for
educational institutions, individuals, companies,
and organizations.
ITSO. International Technical Support
Organization, part of IBM Corp. See
http://www.redbooks.ibm.com for more
information.
Java. Java is a general-purpose concurrent
object-oriented programming language. Its syntax
is similar to C and C++, but it omits many of the
features that make C and C++ complex,
confusing, and unsafe. Java was initially
developed to address the problems of building
software for networked consumer devices. It was
designed to support multiple host architectures
and to allow secure delivery of software
components. To meet these requirements,
compiled Java code had to survive transport
across networks, operate on any client, and
assure the client that it was safe to run. With Java
as the extension language to HTML, a Web
browser is no longer limited to a fixed set of
capabilities. Programmers can write an applet
once and it will run on any machine, anywhere.
JVM . The Java Virtual Machine is the
cornerstone of Sun’s Java programming
language. It is the component of the Java
technology responsible for Java’s cross-platform
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
delivery, the small size of compiled code, and
Java’s ability to protect users from malicious
programs. It is an environment provided for any
conventional operating system (such as UNIX,
OS/2 Warp, or Windows) designed to run Java
applications and applets. (Native Java machines
run Java bytecode directly, without interpretation
or translation.) The JVM knows nothing of the
Java programming language, only of a particular
file format, the class file format. A class file
contains JVM instructions (or bytecodes ) and a
symbol table, as well as other ancillary
information. Java is a trademark of Sun
Microsystems. Consequently a "Java" virtual
machine has passed JavaSoft compliance
testing. See also applet.
MTU. Maximum Transmission Unit, the size of an
IP datagram.
Multicast . Network traffic intended to be received
by more than one system (but not necessarily
every system, as would a broadcast).
NBDD. NetBIOS Datagram Distributor.
NBNS. NetBIOS Name Server. Resolves
NetBIOS names to IP addresses eliminating
typical NetBIOS broadcast traffic.
NC. Network Computer or Network Computing.
NDIS. Network Driver Interface Specification, a
standard used to create drivers for network
adapters.
NetBEUI . NetBIOS Extended User Interface. This
is a non-routable transport protocol written to the
NetBIOS interface. It usually implies NetBIOS
over NetBEUI.
LAN . A Local Area Network is a group of
computers running specialized communications
software, and joined through an external data
path. A LAN will cover a small geographic area,
usually no larger than a single building. The
computers have a direct high-speed connection
between all workstations and servers, and share
hardware resources and data files.
NetBIOS. NetBIOS is a standard programming
interface (API) for the development of distributed
applications. It can run over different transport
protocols, such as NetBEUI, TCP/IP, or IPX/SPX.
Linux . UNIX-compatible operating system
originally developed by Linus Torvalds.
NETBT. Microsoft’s term for NetBIOS over
TCP/IP. IBM calls it TCPBEUI.
LPD. Line Printer Daemon, a server-based
application designed to act as a print server on
TCP/IP networks.
NIC. 1.) Network Interface Card (network
adapter). 2.) Network Information Center. The
most famous one on the Internet is the InterNIC,
which is where new domain names are
registered.
LPR . A TCP/IP-based application which can
forward print jobs to any LPD server.
LSM . IBM LAN Server for Macintosh.
MAC. Media Access Control, pertaining to
network adapters.
MD5 . Message Digest 5. See RSA MD5.
Modem. Modem is an abbreviation for
modulator/demodulator. A modem is a peripheral
device that permits all sort of computers to
receive and transmit data in digital format across
voice-oriented communications links, such as
telephone lines.
MSS . Maximum Segment Size, the maximum size
of a TCP segment.
NFS. Network File System. One of three popular
network file systems, originating with UNIX, to
provide file and disk sharing across a TCP/IP
network. See also AFS and DFS.
NNTP. The Network News Transport Protocol
defines the distribution, inquiry, and retrieval of
news articles on the Internet from TCP/IP sites.
NT. Microsoft operating system (Windows NT).
NTP. Network Time Protocol. It is the protocol
that assures accurate local time keeping with
reference to radio and atomic clocks located on
the Internet.
NVRAM . Nonvolatile Random Access Memory.
Normally, information stored in RAM will be lost
501
without continuous electrical power, but NVRAM
can hold such information intact even without
external power. The IBM Network Station
contains some NVRAM used to store boot
parameters.
NVT. Network Virtual Terminal, as in NVT ASCII.
Defined by Military Standard 1782, which
specifies a standard form for ASCII
representation and for certain other
characteristics of a simple terminal. See also
ASCII.
ODI . Open Datalink Interface, a driver
specification used primarily with Novell networks.
OS/2. Operating System/2. IBM operating system
for the PC.
OSI . Open Systems Interconnection. To support
international standardization of network
terminology and protocols, the International
Standards Organization (ISO) proposed a
reference model of open systems
interconnection.
OSI Reference Model. The OSI model provides
a standard means of describing the data flow in a
network and how it is managed:
7
Application Layer
6
Presentation Layer
5
Session Layer
4
Transport Layer
3
Network Layer
2
Data Link Layer
1
Physical Layer
OSI layer 1 . The Physical Layer. It is the lowest of
the seven defined layers of the generalized
network architecture. It defines the transmission
of bits over a communication channel, ensuring
that 1s and 0s are recognized as such.
OSI layer 2 . The Data Link Layer. It provides
methodologies for transforming the new physical
layer link into a channel that appears free of
errors to the network layer (the next higher layer).
The data link layer accomplishes this by splitting
the input or data stream provided in the physical
layer into data frames that are transmitted
sequentially as messages and by processing the
acknowledgment (ACK) frames sent back over
the channel by the receiver.
502
OSI layer 3 . The Network Layer. It accepts
messages of data frames from the transmitting
host, converts the messages to packets, and
routes the packets to their destination.
OSI layer 4. The Transport Layer. It accepts data
from the session layer (the next layer up, which is
the human user’s interface to the network), splits
this data into smaller units, passes these units
down to the network layer, and ensures that all
the pieces arrive at the destination in correct
order. The transport layer is a true end-to-end
process.
OSI layer 5 . The Session Layer. It is the user’s
interface into the network through which the user
establishes a connection with a process on
another distant machine. Once the connection is
established, the session layer manages the
end-to-end dialog in an orderly manner,
supplementing the application-oriented user
functions to the data units provided by the
transport layer.
OSI layer 6 . The Presentation Layer. Its protocols
format the data to meet the needs of different
computers, terminals, or presentation media in
the user’s end-to-end communications. The
protocols at this layer may also provide data
encryption for security purposes in transmission
over networks, or data compression for efficiency
and economy.
OSI layer 7 . The Application Layer. It specifies
the protocols for the user’s intended interaction
with the distant computer, including such
applications as database access, document
interchange, or financial transactions.
OSPF. Open Shortest Path First routing protocol,
which is the fastest and most efficient routing
protocol. Routers are calculated based on real
performance and are dynamically adjusted if a
route becomes congested. See also RIP.
Packet . A unit of data transmitted at the OSI
network layer; or any addressed segment of data
transmitted on a network.
Peer. A system connected to a network that
communicates more or less on equal terms with
another system on the network, usually providing
basic file and print sharing. Clients, unlike peers,
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
generally must depend on servers for vital
functions. In other words, peers can alternately
behave as both clients and servers in a network.
Requester. A LAN client that requests file and
print services from a server such as OS/2 Warp
Server.
Ping . Packet Internet Groper. A program used to
test accessibility of destinations by sending them
an ICMP echo request and waiting for a reply.
REXX. Restructured EXtended eXecutor
language. Computer programming and scripting
language developed by Mike Cowlishaw of IBM.
Provided as a standard part of most IBM
operating systems. Available for nearly all
operating systems.
PM . Presentation Manager. Graphical subsystem
for displaying windows, icons, and other GUI
components on IBM OS/2 Warp operating
system.
POP. The Post Office Protocol is one of mail
exchange protocols.
PPP. Point-to-Point Protocol. The successor of
the SLIP protocol, PPP allows a computer to use
a regular telephone line and a modem to make IP
connections. PPP can also carry other routable
protocols such as IPX.
Protocol. The rules by which two network
devices trade information in order to
communicate. Must include rules about a lot of
mundane detail, as well as rules about how to
recover from a lot of unusual communication
problems. Thus, the rules can be quite
complicated.
Proxy. An agent which can forward requests to
another part of the network on behalf of clients,
returning any results. Most often used to allow
Web browsers to access public Web sites beyond
a firewall.
PTF. Program Temporary Fix, issued by IBM to
correct a defect in an IBM software product.
QOS. Quality Of Service. Typically a specified
bandwidth or timeliness of delivery of packets.
Controlled load reservation has a specified data
rate, whereas guaranteed reservation additionally
has an upper bound on delivery time.
RARP. The Reverse Address Resolution Protocol
is used to map the MAC, or hardware address, to
a host’s IP, or software address.
RedHat . One of two major vendors supplying
commercial versions of Linux. See also
Slackware.
RFC. Request for Comments. Refers to
documents published by the Internet Engineering
Task Force (IETF) that have achieved at least
some degree of standardization. RFCs describe
various TCP/IP protocol standards.
RIP. Routing Information Protocol. This protocol
allows routers to exchange routing details on a
network. Using RIP, routers can create and
maintain a database, or routing table, of current
information. Workstations can query the nearest
router to determine the fastest route to a distant
network by broadcasting a RIP request package.
Routers send periodic RIP broadcast packets
with current information to keep all routers on the
internetwork synchronized.
RIPL . Remote Initial Program Load. Method of
obtaining boot software across a network to start
a PC or other computer, often one without a disk
drive.
Router. A network relay that uses a protocol
beyond the datalink protocol to route traffic
between LANs and other network links.
Routing Protocol. A protocol sent between
routers by which routers exchange information on
how to route to various parts of the network.
TCP/IP includes several routing protocols such
as RIP, EGP, BGP, OSPF, and dual IS-IS.
RPC. Remote Procedure Call.
RSA MD5. Rivest-Shamir-Aleman algorithm
Message Digest 5.
RSVP. Resource ReSerVation Protocol. A
protocol to reserve quality of service (QOS) for a
data stream. See 10.5.2, “RSVP” on page 408.
SCSI. Small Computer Systems Interface. A
high-speed interface bus used for disk drives,
503
tape drives, scanners, printers, CD-ROM drives,
digital cameras, and other devices. Available in
several versions including SCSI-I, SCSI-II (Fast
SCSI), Wide (16-bit data path) and UltraWide.
Segment. Effectively a TCP datagram.
ServicePaks . See CSD.
Slackware. One of two major vendors supplying
commercial versions of Linux. See also RedHat.
SLIP. Serial Line (or Link) Internet Protocol.
Method used to provide a TCP/IP connection
over serial connections, such as dial-up modems.
See also PPP.
SMTP. Simple Mail Transport Protocol, designed
to send e-mail across a TCP/IP network.
SNMP. The Simple Network Management
Protocol is one of the most comprehensive tools
available for TCP/IP network management. It
operates through conversations between SNMP
agents and management systems. The SNMP
management system can collect statistics from
and modify configuration parameters on agents.
SOCKS. A technique used on a TCP/IP network
to provide secure access through a firewall for
most TCP/IP-based applications.
Solaris . Sun’s version of the UNIX operating
system.
Spam. Unsolicited e-mail, usually sent to huge
numbers of people. Also a brand of canned meat,
trademark of Hormel.
SPX . A network protocol created by Novell for
use with NetWare products.
SrvIFS. A simple file system created by IBM.
Used primarily to distribute software to PCs
booted from OS/2 Warp diskettes. See also IFS.
Stack . The collection of drivers, APIs, and other
files used to implement a network protocol on a
computer, as in a TCP/IP protocol stack.
Subnet. The primary reason to divide a network
into subnets is network performance and
available bandwidth. Without separate networks,
each transmission would be broadcast across the
entire internetwork, waiting for the destination
504
system to respond. Routers divide, as well as
provide communications between the networks.
Switch . A network device used to alter or change
network connections, often to help boost
performance.
TCP. The Transmission Control Protocol defines
connection-oriented, or reliable, transport. See
also UDP.
TCPBEUI. IBM NetBIOS over TCP/IP
implementation.
TCP/IP. Transmission Control Protocol/Internet
Protocol. The Internet is based on TCP/IP
network protocols.
TFTP. The Trivial File Transfer Protocol is part of
the TCP/IP protocol suite that is used to transfer
files between any two computers, provided they
support TFTP. Unlike FTP, TFTP does not use an
user ID to ensure proper access rights. However,
a password can be set optionally in many TCP/IP
implementations.
Token-Ring . People often use the term
"Token-Ring" to designate IEEE 802.5. In the
more general sense of the term, a token ring is a
type of LAN that has stations wired in a ring,
where each station constantly passes a special
message (a token) on to the next. Whichever
system has the token can send a message.
UBE . Unsolicited Bulk e-mail. (See Spam)
UCE . Unsolicited Commercial e-mail. (See
Spam)
UDP. The User Datagram Protocol is a transport
protocol in the Internet suite of protocols. It
provides a connectionless, or unreliable,
transport.
UNC. Universal Naming Convention. A method
used to name network resources on
NetBIOS-based file and print sharing systems,
such as Warp Server.
UNIX . Operating system originally developed by
AT&T’s Bell Labs. Historically one of the most
popular operating systems used to develop
Internet and TCP/IP technologies.
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
URL . The Universal Resource Locator is the
pathname of a document on the Internet. URLs
can be absolute or relative. An absolute URL
consists of a prefix denoting a method (http for
Web sites, gopher for gophers, ftp for FTP sites,
and so forth). The prefix is followed by a colon
and two slashes (://), and an address. The
address consists of a domain name followed by a
slash and a pathname (or username @domain
name for mailto). The last part is optional.
505
506
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
Index
A
A record
record updates 23
AIX 40, 102, 277, 371, 380, 406
Configuring DHCP server 50
DDNS 218
DHCP and BOOTP 62
DHCP client options missing 89
dhcpsconf 50
DNS 211
DNS client 209
DNS server 208
Firewall 362
HACMP 380
Name serving 208
Start DHCP server (smit dhcpsd) 63
X-windows 51
AIX commands
nsupdate 222, 224
generate zone key 223
AIX DDNS
controlled mode 222
KEY resource record 223
AIX files
dhcpaction 224
named.conf 222
named.forward 223
named.rev 224
alias names 13
Apple Macintosh 40, 77
ARP 387
authoritative 17
authoritative name server 20, 21
B
BIND 150, 178
BIND 8.1.1 157, 159
B-Node 232, 233
enhancing DOS, Windows 3.1, and Windows for
Workgroups 233
enhancing OS/2 Warp 239
enhancing Windows 95/NT clients 236
BOOTP 10, 79, 85, 86, 90, 91, 110, 117, 285, 361,
367, 457
BOOTP relay agent 285
© Copyright IBM Corp. 1998, 2000
BOUND 114
broadcast 232, 241, 291, 408
C
caching-only name server 20
CIFS 252
Class A 3, 6, 8
Class B 4, 6, 8, 9
Class C 4, 7, 8, 9, 44
Class D 4
Class Determination 5
Class E 4
CNAME
resource record 24
Commands
APPLET.EXE 424
APPLETVIEWER.EXE 424
ARP 387
control-panel 90
DADMGUI.CMD 412
DADMIN 41, 85, 193, 379
DDNSAPC.CMD 412
DDNSAPS.CMD 420
DDNSCFG.EXE 166, 172, 199, 200, 204, 207,
208
DDNSZONE 178
DETACH 108
DHCPMON 68, 70
DHCPRD.EXE 289
dhcpsconf 52
DHCPSCPC.CMD 412
DHCPSCPS.CMD 420
DHCPSD.EXE 378
grep 62
HOST 177, 388, 397
IBMDYNIP.EXE 143, 145, 147
ifconfig 92, 100, 405
IPCONFIG 74, 76, 145
IPFORMAT 386
IPGATE 279, 362
IPREPORT 387
IPTRACE 386
kill 63
LPR 110
MAPNAME.EXE 243
NBTSTAT 251
NBUTIL 233, 235
507
NETSTAT 195, 387, 395, 396, 405
NSLOOKUP 177, 389, 397, 399
PING 99, 233, 384, 393
refresh 63
REGEDIT2 344
RFCADDR.EXE 242, 331
route 101, 280, 283, 406
SEDISK 430
SETBOOT 377
shutdown 92
smit 63
startsrc 63
startx 90
stopsrc 63
TCPCFG 69, 201
TCPCFG2 67, 80, 161, 198, 204
TEDIT 430
THINLAPS 430
touch 101
TRACERTE 385, 395, 405
WINIPCFG 72, 147, 148
computername 231, 358
D
DADMIN 41
Daemons
DHCPRD 290
FTPD 34, 372
INETD 86, 373
LPD 108
LPRPORTD 108
portmapper 86
REXECD 372
ROUTED 281
syslog 80
TELNETD 34
TFTP 86
datagram 2
datagram distributor 256, 262
datagrams
NetBIOS 241
DB2 445
UDB Version 6.1 445
DDNS 104, 150, 319, 348
AIX 218
AIX pre-requisites 208
Dynamic Presecured Domain (in OS/2 Warp 4)
203
508
Dynamic Presecured Domain (in Windows 95)
147
Dynamic Presecured Mode 154, 168, 350
Dynamic Secured Mode 154, 162, 350
ns-updates 166, 172
ProxyArec 155, 159, 176, 356
PTR records 168, 175
Windows client 131
WINS 131
DDNS Parameter
DDNSAdministratorClient 183
deferUpdCnt 183
incrTime 183
keyToSec 183
notify 180
notify.add 180
notify.delayTime 180
notify.remove 180
notify.retryNumber 180
notify.retryTime 180
reverseMapping 183
safeWrite 181
sepDynStatic 182
sigDEL 181
sigDEL.time 181
timeSync 181
timeSync.toSecondaries 181
ttlSet 182
ttlSet.value 182
delegation 16
DHCP 29, 42, 90, 91, 285, 319
Basics 113
Client states 113
Lease
Eviction 116
Rebinding 115
Renewing 115
Messages 118
DHCPDISCOVER 118
DHCPOFFER 120
DHCPREQUEST 122
see also dynamic IP
States
BOUND 114
INIT 114
INIT-REBOOT 114
REBINDING 114
RENEWING 114
SELECTING 114
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
DHCP Boot Diskettes 429
DHCP Clients 64
AIX 86
options missing 89
Apple Macintosh 77
DOS 92
Hewlett-Packard Laser Jet 4000 105
IBM Network Station 79
IBM WorkSpace On-Demand 102
Linux 90
OS/2 Warp 4 64
Sun 99
Windows 2000 75
Windows 3.1 92
Windows 95 72
Windows 98 72
Windows for Workgroups 92
Windows NT Workstation 73
DHCP Monitor 68, 201, 203
DHCP Options 136
Option 001 49
Option 003 45
Option 006 45, 136, 166, 174
Option 009 328
Option 012 358
Option 015 37, 45, 136, 166, 174
Option 044 140, 262, 327
Option 045 262, 327
Option 046 140, 262, 327
Option 069 190
Option 070 191
Option 071 328
Option 077 360
Option 081 358
Option 114 334, 335
Option 150 325, 329
Option 151 325, 329
Option 192 24, 352
Option 200 328
Option 200-208 324
Option 201 328
Option 202 328
Option 205 324, 341, 342
Options 001-208 457
DHCP relay agent 285, 360
DHCP Server 64
DHCP Servers
AIX 50
OS/2 Warp Server for e-business 30
Windows NT 43
DHCPCD.INI 90
DHCPDISCOVER 118, 285
DHCPOFFER 120
DHCPREQUEST 122
dhcpsd
subsystem 222
digital signature 152, 153
DLS 92, 97, 233
DNS 13, 18, 19, 42
AIX 211
AIX client 208, 209
AIX server 208
AIX server configuration 211
name services 12
name system 12
primary 160
servers 12
static 13
translating 12
Windows client 131
Windows NT 131
Windows NT limitations 131
DNS name space 14, 21
domain 14, 16, 20, 231
child 21
parent 21
primary 160, 163, 169
reverse 134
domain name
fully qualified name 13
domain name system 16
DOMAINSCOPE 243
Domino Go 411
Domino Go Webserver 411, 420, 426
DOS 40, 92
DOS LAN Services 92
dynamic addres
assign 11
dynamic IP 29
BOOTP Clients 84
DHCP 29
DHCP Clients 64, 67, 72, 73, 75
DHCP Manager 45
DHCP Options 44, 47, 82
see also DHCP Options
DHCP Server 35, 84, 102
DHCP server 30, 44
DHCPDISCOVER 285
509
Dynamic IP Client
95/98/NT pre-requisites 143
E
encryption
key
public 223
Ethernet 91, 100
F
Feature Installer 31, 64, 66
Files
AUTOEXEC.BAT 235, 295, 443
cache.dns 140
CONFIG.SYS 34, 262, 430
DDNS.DAT 153, 205
dhcp.hme0 101
DHCPCD.CFG 207, 263, 326, 328, 334, 359
DHCPRD.CFG 289
DHCPS.AR 378
DHCPS.CR 378
DHCPSD.CFG 36, 356
dhcpsd.cnf 50, 60, 61
DHCPSD.LOG 39
DNSEXT.CFG 178
DNSF0000.DOM 184
DNSF0000.STA 184
hme0.dhc 101
hostname.hme0 101
HOSTS 42, 162, 166, 169, 254
HTTPD.CNF 420
IBMLAN.INI 241, 332
IBMTOKCS.NIF 430
LANTRAN.LOG 438
LMHOSTS 236, 254
masterfile.cnf 59
MPTSTART.CMD 31
NAMED.BOOT 184
NAMED.BT 184
NBUTIL.LST 235
NETSCAPE.INI 341, 344, 345
NETWORK.INI 97, 442
NSUPDATE.LOG 177
NTS-SRVR.CFG 295, 298
NTS-SRVR.DHC 298, 299
NTS-SRVR.DNS 298, 299
NTS-SRVR.NBN 298, 299
NTS-SRVR.RUR 299
510
PROTOCOL.INI 243, 257, 262, 434
RESOLV2 398
RFCBCST.LST 241, 331, 334
RFCNAMES.LST 239, 331
RHOSTS 42
SENDMAIL.CF 186, 187, 188
SENDMAIL.UML 189
SETUP.CMD 31, 34, 279, 280, 362
SOCKS.ENV 334, 340
STARTUP.CMD 331
TCPSTART.CMD 31, 282
TFTPAUTH 364
TFTPPAUTH 83
WCLLOCAL.INI 321
firewall 22, 25, 341, 361
Firewalls
protection 25
proxy and SOCKS 25
VPN 25
FixPak 30, 344
forward mapping 132
forwarder 22
FQDN 355, 359, 420
fully qualified domain name 12
G
gateway 9
groupname 231
H
HACMP 380
Hewlett-Packard 41, 105
HINFO
resource record 24
see also name server
H-Node 232, 261
HOST 388, 397
host 8
host name 12
hostname 34, 53, 134
I
IBM Network Station 40, 79, 85
IBM OS/2 NetBIOS OVER TCP/IP 240
IBMWARP_V3.1 326, 327
IEEE 802.2 data link protocol 227
INIT 114
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
INIT-REBOOT 114
InJoy
IP masquerading 292
internal root 21
Internet 22, 150, 284
root server 15
intranet 22, 25, 361
IP address 10, 15, 42, 150, 166, 169, 174, 177,
242, 360
IP datagram
MTU 27
IP masquerading 291
IP packet 293
IPTRACE 386, 387
J
Java 31, 35, 64, 66, 340, 412
JDK 424, 425
JetDirect 105
JIT 412
K
KEY
AIX DDNS resource records 223
resource record 24
see also name server
Keyfile 24
L
LAN 27
LAN interface 0 33, 67, 69
LAN Requester 332
lease 41, 46, 68, 360, 381
lease time
choosing the right one 381
examples 382
multiple leases 383
Rule-of-Thumb 382
Linux 40, 90
LMHOSTS 237
localhost 34, 420
lookup 241
loopback 5, 34
Lotus Domino 426
M
MAC 82, 353, 356, 360
mail
routing 15
media
AIX 208
M-Node 232
MPTS 31, 66, 240, 241, 260
MSBROWSE 231
MTU 26, 404, 405
MX
resource record 23
see also name server
N
name server 16, 29
authoritative 20
caching-only 20
child 20
dynamic 22
firewall 22
forwarder 22
master 19
non-authoritative 20
parent 20
parent and child 20
primary 19
record types 23
A 24, 150, 151, 165, 171, 246, 348, 356,
400
CNAME 24, 185, 245, 247, 355, 400
HINFO 24, 245, 400
KEY 24, 153, 154, 185, 349, 400
MB 400
MINFO 400
MX 23, 185, 186, 193, 400
NS 23, 141, 185, 400
PTR 24, 135, 150, 151, 158, 160, 165, 167,
171, 174, 353, 356, 400, 447
RP 24
RR 24, 348, 356
SIG 349, 401
SOA 23, 141, 185, 401
TXT 24, 184, 401
UINFO 401
WKS 401
root 21
secondary 19
zone transfer 20
Name serving
511
AIX 208
name space 16, 136
named
subsystem 222
NAT
IP masquerading 293
NBNS 27, 51
Configuring DOS, Windows 3.1, and Windows
for Workgroups 257
Configuring OS/2 Warp 260
Configuring Windows 95 258
Configuring Windows NT 259
see NetBIOS name server 232
NC 86
Neighborhood Browser Enabler 271, 272
NET commands
NET START 235
NET START BROWSER 274
NET START DHCPSERVER 45
NET START DNS 133
NET START WINS 137
NET STOP BROWSER 274
NET STOP SERVER 273
NET.QUESTION 445
NETAnywhere 319
NetBIOS name resolution 246
NetBIOS name server 227, 323, 327
Netfinity Client 448
NetFinity Manager 390
NetFinity Server 447
NetFinity Service Manager 448
Netscape 31, 64, 66, 196, 334, 418
NetScape Navigator 418
Netscape Navigator 451
NETSTAT 387, 395, 396
Network Address Translation 292
NFS and Windows NT 85
Node Status request 246, 249
Non-RFC-encoded names 246, 252
NS
resource record 23
see also name server
NSLOOKUP 350, 379, 389, 397, 399
NSUPDATE 144, 146, 148, 151, 348
ns-updates 165, 172
NT
see also Windows NT 43
NVRAM 86
512
O
OS/2 Warp 4 40, 49, 64, 67, 198, 201, 203, 323,
335, 429
DHCP relay agent 289
NBNS Parameter
NBDD 261
NBNS 261
router enablement 279
OS/2 Warp Server 102, 246, 256, 277, 319
Configuring DDNS Dynamic Presecured Mode
169
Configuring DDNS Dynamic Secured Mode
162
Configuring DHCP 35
DDNS Security 153
DDNS updates requests 154
DHCP Clients 64
DHCP server 30
Dynamic DNS (DDNS) 150
Dynamic DNS Clients 198
Dynamic Presecured Mode 159
Dynamic Secured Mode 159
Neighborhood Browser Enabler 256
NSUPDATE 151
ns-updates 166, 172
ProxyArec 155, 176
Remote IPL 207
Routing 33
TCP/IP Administrator Password 32
OS/2 Warp Server for e-business
Installing DDNS server 161
Installing DHCP server 30
Neighborhood Browser Enabler 274
OS2POPS 186, 196
P
parent and child name server 20
password for OS/2 TCP/IP administrator 32
Peer Services 93
PING 384, 405
P-Node 232, 261
PPP clients 321
PPP dialer 319
PPP Dial-Up 317
PPP Server 318
OS/2 Warp Server 318
primary name server 19
private network 22
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
protect mode redirector 96
Protocols
ARP 2, 27, 387, 388, 393
BOOTP 79, 84
FTP 26, 362
HTTP 26, 362
ICMP 2, 26, 384, 388
IEEE 802.2 27
IGMP 2, 27, 388
IMAP 26
IP 2, 26
IPv6 26
IPX 27
NBT 227
NetBEUI 27, 93, 227, 229, 408
NetBIOS 27, 93, 227, 229
NetBIOS over TCP/IP 227
NNTP 26
non-routable 227
OSPF 277
POP 26
POP3 186
PPP 27, 317
RARP 2, 27, 105
RIP 277, 281, 284
routable 227
RSVP 408
SLIP 27
SMB 247, 252
SMTP 26, 186, 188, 195
SNMP 26
TCP 2
TCP/IP 227
TCPBEUI 93, 227, 229, 239, 242, 323, 408
sockets-send commands 228
translation 228
TFTP 80, 81, 361
UDP 2, 227, 388
UDP/IP 227
proxy
IP masquerading 293
ProxyArec 23, 155, 356, 358
dial-in 23
Option 81 considerations 358
Protected 358
Standard 358
PTR
record update 22
resource record 24
see also name server
public key 153
public-private key 156
R
REBINDING 114
RedHat Linux 90
Redirector 96
Relay Agents 118
Remote administration 411
Remote IPL 102
RENEWING 114
resource record
A 24
reverse mapping 132, 134, 171
REXX 104
REXX Commands
DHCPIBM.CMD 329
FAILSAFE.CMD 375
GETNBNS.CMD 263, 266, 271
MYOFFICE.CMD 329
MYWEBMSG.CMD 335, 336
NSSOCKS.CMD 341, 342
RDADMGUI.CMD 415
RDDNSAPC.CMD 413
RDHCPSCP.CMD 416
SEEMYMSG.CMD 339
SETNBNS.CMD 263, 266, 270
RFC 27
RFC-encoded name 246, 249
root name server 21
router 9, 29, 277
routing table 10, 278
RP
resource record 24
see also name server
RR
resource record 24
see also name server
RSA 150, 153, 154, 353
1024-bit key 349
128-bit key 349
56-bit key 349
public/private key pair 348
RSVP 408
S
Scope 46, 49
513
secondary name server 19
Security 347
TFTP 364
SELECTING 114
Service Pack 5 43
Shadow IPserver 255, 356, 380
Installing and Configuring 295
OUTAHERE parameter 297
SNA 408, 409
Sniffer 117, 119, 121, 123, 125, 127, 129
SOA
resource record 23
see also name server
sockets 26
SOCKS 25, 34, 324, 334, 340, 341, 362
socks
IP masquerading 293
Software Choice 30, 64, 272, 318, 418
subdomain 14, 16, 17, 20
subnet 6, 37, 174, 242, 277, 291
subnet mask 7, 29, 49
subnetting 9
subsystem
dhcpsd 222
named 222
suffix
0x00 244
0x03 244
0x20 244
Sun JDK 418
Sun Microsystems 99
Sun Microsystems Solaris 40
Supernetting 9
T
TCP/IP 31
setting Admin password in OS/2 32
TCP/IP Version 4.21 64
Telnet 105, 106
TFTP and Windows NT 85
Tivoli 446
TMA 446
TME 10 446
TMR Server 447
Tivoli Managed Agent 446
TMA 446
token-ring 91, 100
TRACERTE 385, 395
514
Troubleshooting 390
T-Sig 157, 159
TTL 253, 349
TXT
resource record 24
see also name server
U
unique domain name 15
UNIX 398
untrusted network 22
URL 334
translation 12
V
virtual redirector 96
VPN 25, 363
W
WAN 381, 408
Warp Server
See OS/2 Warp Server
Web Server 418
Windows 2000 40, 75
Windows 3.1 40, 92, 233
Windows 95 40, 49, 72, 143, 150, 196, 236, 358
Modifying the registry 149
Windows 95/98
DNS client configuration 142
Windows 98 40, 72, 236
Windows for Workgroups 40, 92, 233
Windows NT 40, 49, 73, 75, 85, 102, 131, 246,
358
Configuring DHCP 45
DHCP relay agent 286
DHCP server 131
DNS (static) 131
DNS client configuration 142
Firewall 362
Installing DHCP 43
Installing DNS 132
Node Status request 247
router enablement 282
WINS 131, 136
WINS 131, 227, 231, 248, 253, 258
configuration 137
DNS integration 136
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
installation 137
Windows Internet Naming Service 131
workload 19
WorkSpace On-Demand 40, 102, 206, 358, 362
Z
zone 19, 132, 350
primary 134
zone key 165
zone of authority 16
zone transfer 19, 20
515
516
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
IBM Redbooks evaluation
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
SG24-5280-01
Your feedback is very important to help us maintain the quality of IBM Redbooks. Please complete this
questionnaire and return it using one of the following methods:
• Use the online evaluation form found at http://www.redbooks.ibm.com/
• Fax this form to: USA International Access Code + 1 914 432 8264
• Send your comments in an Internet note to [email protected]
Which of the following best describes you?
_ Customer _ Business Partner
_ Solution Developer
_ None of the above
_ IBM employee
Please rate your overall satisfaction with this book using the scale:
(1 = very good, 2 = good, 3 = average, 4 = poor, 5 = very poor)
Overall Satisfaction
__________
Please answer the following questions:
Was this redbook published in time for your needs?
Yes___ No___
If no, please explain:
What other Redbooks would you like to see published?
Comments/Suggestions:
© Copyright IBM Corp. 1998, 2000
(THANK YOU FOR YOUR FEEDBACK!)
517
SG24-5280-01
®
Beyond DHCP - Work Your TCP/IP Internetwork with Dynamic IP
SG24-5280-01
Printed in the U.S.A.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement