Detecting and Preventing Rogue Devices on the

Interested in learning
more about security?
SANS Institute
InfoSec Reading Room
This paper is from the SANS Institute Reading Room site. Reposting is not permitted without express written permission.
Detecting and Preventing Rogue Devices on the
Network
AD
Copyright SANS Institute
Author Retains Full Rights
ut
ho
rr
eta
ins
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
07
,A
Detecting and Preventing Rogue Devices on the Network
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
20
GCIA Gold Certification
tu
te
Author: Ibrahim Halil Saruhan, ibrahimsaruhan@gmail.com
Accepted: August 8, 2007
©
SA
NS
In
sti
Adviser: John Bambenek
Ibrahim Halil Saruhan
© SANS Institute 2007,
1
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
Table of Contents
Abstract ............................................................................................................................. 5
2.
Introduc ti on .................................................................................................................... 5
3.
Architecture .................................................................................................................... 9
fu
ll r
igh
ts.
1.
AP Architecture ....................................................................................................................... 9
eta
ins
Client Architecture ...............................................................................................................10
ho
rr
Server Architecture..............................................................................................................10
07
,A
ut
Zone Architecture.................................................................................................................10
Location ....................................................................................................................................10
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Assignment ..............................................................................................................................11
Set-Up................................................................................................................................11
5.
Monit or ing ......................................................................................................................13
NS
In
sti
tu
4.
SA
CPE Site Survey.....................................................................................................................14
©
CAU Site Survey ....................................................................................................................16
6.
Detecti on .........................................................................................................................19
Man in the Middle Attack ..................................................................................................20
Ibrahim Halil Saruhan
© SANS Institute 2007,
2
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
Evil Twin Attack .....................................................................................................................27
MAC Address List Match...................................................................................................28
fu
ll r
igh
ts.
Client MAC Address Match ..........................................................................................28
AP MAC Address Match ................................................................................................29
ins
MAC Address Spoofing .....................................................................................................30
eta
RF Jamming.............................................................................................................................31
ho
rr
DOS Attacks ...........................................................................................................................33
07
,A
ut
Duration Attack .................................................................................................................34
20
Deauthentication ...............................................................................................................35
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
tu
te
801.1x EAP Authentication ..............................................................................................38
Preventi on ......................................................................................................................39
8.
C onclusion ......................................................................................................................44
9.
Future Work ..................................................................................................................45
©
SA
NS
In
sti
7.
10 . References .....................................................................................................................47
11 . Appendix ..........................................................................................................................50
Source Code ...........................................................................................................................50
Ibrahim Halil Saruhan
© SANS Institute 2007,
3
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
Client MAC Address List Check .................................................................................50
Client MAC Address Spoofing Check......................................................................50
fu
ll r
igh
ts.
AP MAC Address List Check .......................................................................................51
Main Monitor.......................................................................................................................52
ins
Log Traffic...........................................................................................................................53
eta
RF Jamming.........................................................................................................................54
ho
rr
Client Site Survey .............................................................................................................54
07
,A
ut
AP Site Survey ...................................................................................................................57
20
Man in the Middle Attack ..............................................................................................59
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
©
SA
NS
In
sti
tu
te
Evil Twin ................................................................................................................................60
Ibrahim Halil Saruhan
© SANS Institute 2007,
4
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
1.
Abs tract
Rogue device detection is an important aspect in wireless security. Without
proper detection analyzing could be difficult and error-prone. There are known
fu
ll r
igh
ts.
techniques for rogue detection, nevertheless this paper will try to solve the
problem from a different perspective. The solution in this paper can be applied to
all wireless networks.
ins
The main approach of this paper is to show how to use site survey to detect
eta
rogue devices in a wireless network. Site survey, if used correctly is extremely
rr
beneficial for detecting rogue devices. Rogue device detection can be considered
ho
the initial phase of wireless intrusion detection, in case it is not feasible to install
In troduction
Key
fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
20
2.
07
,A
ut
sensors to cover all the wireless network area.
tu
te
Rogue access point detection is an important aspect of wireless IDS. (Potter,
sti
2007) Rogue access points can pose significant threats. The first step in dealing
In
with rogue devices is to find out that they exist. A radio device somewhere must
SA
NS
note the existence of an unauthorized device. (Gast, 2005)
In this paper rogue device detection is implemented by using different
©
techniques like site survey, MAC address list checking, noise checking and
eventually wireless traffic analysis. I created some strategies by using IDS
methodologies to detect rogue devices interloping on the network. The
experiments are done on a broad wireless network covering a few square miles.
The coverage area could have been bigger in theory, but the sample was enough
Ibrahim Halil Saruhan
© SANS Institute 2007,
5
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
to do analysis.
Rogue access points come in two varieties. Internal rogue access points are
those that (for instance) an employee brings in and plugs into a corporate network.
fu
ll r
igh
ts.
The access point is outside the control of IT personnel and serves as a gateway for
attackers to enter the enterprise. The other type of rogue access point is more
difficult to control. The external rogue access point is one that is controlled by an
attacker and designed to spoof legitimate clients into connecting to it rather than
ins
the correct access point. Usually this is accomplished by setting the rogue access
eta
point SSID to the same SSID as the friendly access point and then boosting the
rr
signal of the rogue access point. This will cause client associations to come to the
ut
ho
rogue access point. The attacker may then attempt to steal userʼs credentials via
07
,A
spoofed web pages and portals designed to trick users into giving up passwords,
20
credit card numbers and other personal information. These types of rogue access
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
points are generally easy to detect but difficult to turn off as the attacker then
tu
te
needs to be physically located. (Potter, 2007)
In
sti
Rogue can be defined in many other ways. Some say anything other than
NS
legitimate clients and APʼs are rogue devices. Another approach is that anything
SA
on the wireless networks that isnʼt authenticated is rogue. For this paper, rogue is a
device, AP, or client which is trying to connect attack or interfere to the wireless
©
network. Briefly in this paper, rogue is accepted as the malicious entity.
I did most of the testing on dedicated APʼs which are not connected to the
wired networks. The purpose of an AP in this paper is basically a device which
provides service to its clients to connect to the internet. An AP can connect to the
Ibrahim Halil Saruhan
© SANS Institute 2007,
6
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
internet by using different methods like wired LAN by using Ethernet, dedicated T1
lines or some other means of connection, such as 802.11a and WiMAX.
APʼs can be installed on towers or site hosts. In my testing set-up, APʼs are
fu
ll r
igh
ts.
installed on site hosts. Every client has a CPE (Customer Premises Equipment)
installed at their office or home. CPE's connect client computers to the APʼs. There
could be instances like someone spoofing MAC address of CPEʼs or APʼs. On the
other hand someone can create noise intentionally like applying broadcast de-
eta
ins
authentication attack to one of the APʼs or unintentionally.
rr
Moreover someone can create DDOS attack to the AP. At the same time
ho
they can start broadcasting with the same SSID. In this case they can ensure AP is
ut
not functioning anymore and at the same time spoof the MAC Address of the AP
07
,A
and show clients a fake login page, which is called Evil Twin attack. The point in the
20
Evil Twin
is =
the
rogue
spoofʼs
MAC
address
of the
legitimate AP
Key attack
fingerprint
AF19
FA27device
2F94 998D
FDB5the
DE3D
F8B5
06E4 A169
4E46
te
and creates an evil twin. It also has to create DOS attack on the legitimate AP and
tu
broadcast with higher power. There might also be cases where someone can start
In
sti
broadcasting with the same SSID but on a different channel and only for listening
NS
purposes, which is called Monkey-in-the-middle attack or Man-in-the-middle attack.
Most of these cases are tested in this paper. There is some information for each of
©
SA
these cases and also detection strategies.
On the other hand, there could be cases where an AP can broadcast with
higher power and same SSID. MAC Address spoofing is also one of the cases that
need to be checked. APʼs that I was testing are working 24X7. Some enterprises
turn off the access points during off hours to prevent war-driving efforts, but this
Ibrahim Halil Saruhan
© SANS Institute 2007,
7
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
is not the case for this paper.
A rogue AP can be both hardware and software. There are soft APʼs like
HOSTAP or FakeAP; nonetheless I decided to use a wireless USB adapter. I bought
fu
ll r
igh
ts.
a brand new USB wireless finder and adapter Zyxel 225-AG. It also functions as an
AP. This functionality works only in XP.
Once an AP is discovered, the next step is to identify whether it is a rogue
ins
AP or not. One way to do this is to use pre-configured authorized list of APs. Any
eta
newly detected AP that falls outside the authorized list would be tagged rogue.
rr
Some of the different ways in which IT managers can populate the authorized list
ut
ho
are:
07
,A
• Authorized MAC
sti
tu
• Authorized Vendor
te
20
Key
fingerprint =SSID
AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
• Authorized
NS
In
• Authorized Media Type
SA
• Authorized Channel (Anand, 2004)
©
In this paper MAC address and SSID are checked. Vendor type, media type
and channel are not used to detect rogue APʼs.
Detection components may be implemented through the use of a scanning
feature that searches periodically for unauthorized devices or dedicated scanning
devices. (Gast, 2005)
Ibrahim Halil Saruhan
© SANS Institute 2007,
8
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
The main approach in this paper is using client devices to do periodical
scanning instead of using dedicated scanning devices. Real-time detection of a
rogue device on the wireless network depends on continuous monitoring. Because
fu
ll r
igh
ts.
clients are up almost all the time, doing site survey on clients and collecting data in
a central server was adequate.
3.
Architecture
eta
ins
AP Architecture
rr
There are different kinds of equipment that can be used as an AP. To do the
ho
testing, I used specifically central access units (CAU) which supports 802.11b
ut
protocol. Each CAU is composed of three independent single board computers.
07
,A
One of them is the main single board computer that does most of the
tu
te
20
functionality.
The other
are
secondary
single
board
computers
which act only
Key fingerprint
= AF19two
FA27
2F94
998D FDB5
DE3D
F8B5 06E4
A169 4E46
as a wireless-to-wired bridge. Each single board computer functions as an AP.
sti
Wireless radio cards on all the single board computers are identical. They
In
support 802.11b protocol use Prism 2.5 chipset and have 23 dBm RF output
NS
power. They also support wireless extensions. Each CAU connect to the internet
SA
through T1, multiple T1, or other solutions. On each single board computer, hostap
©
is running to get the air traffic and forward it to the main single board computer.
802.1x is used for authentication.
Perl scripts, which send http requests to CPE and parse the http responses,
run in APʼs.
Ibrahim Halil Saruhan
© SANS Institute 2007,
9
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
Client Architecture
Customer premises equipment (CPE) is a device which includes a wireless
card and an antenna. It is connected to the clientʼs computer. Each CPE has a
fu
ll r
igh
ts.
Prism 2.5 chipset wireless card and also runs a small web server to respond http
requests to do site survey. The CPEʼs entire purpose is to connect the client
computer to the AP.
ins
Server Architecture
eta
A debian based Linux server is used to do logging. Apache server is installed
ho
rr
in the server. The main project is developed in PHP by using a 2-tier application
ut
model for simplicity. It stays in the Linux server. SSH2 is installed in this server and
07
,A
used to send data to the APʼs securely. There are also cron jobs running on the
20
server.
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
tu
te
Z one Architecture
sti
Every zone constitutes a CAU (3 APʼs) and multiple CPEʼs connecting to it. In
In
a zip area there could be multiple zones. For example the first zone in 84095 zip
©
Loca t i on
SA
NS
area would be called 8409501 and the second zone would be called 8409502.
Wireless access points can lose signals because of walls, doors, floors,
insulation and other building materials. The signals may also enter into another
userʼs airspace and connect with their wireless local area network. This is referred
to as accidental associations and can occur in densely populated areas where
Ibrahim Halil Saruhan
© SANS Institute 2007,
10
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
several people or businesses use wireless technology. (Lane, 2005)
An APʼs placement and signal strength have to be calibrated or blocked to
make sure the transmitting coverage is just enough to cover the correct area. The
fu
ll r
igh
ts.
RSSI (Received Signal Strength Indicator) on a wireless card is a good way of
measuring wireless coverage inside and outside of a WLAN perimeter. (Hutchison,
2004)
ins
The signal strength needed to make a connection is much higher than that
eta
needed to just listen into the network traffic. So by its nature itʼs a lot easier to
ho
rr
just listen than it is to make a legitimate connection. (Hutchison, 2004)
ut
Based on all these, the location of APʼs and their connecting clients are
07
,A
selected accordingly.
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Assignment
tu
te
Assignment of clients to the APʼs is done based on physical location and
In
NS
Set-Up
SA
4.
sti
signal strength.
©
This project has multiple entities which communicate to each other. Clients
communicate with APʼs and APʼs communicate with the central server. Clients
transfer data to AP via http response after the http request from AP and AP
transfers the data to the central server via XML. Central server parses the XML
data and logs it in a MYSQL database table for further analysis. There are cron
Ibrahim Halil Saruhan
© SANS Institute 2007,
11
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
jobs running, which check the database periodically and as soon as there is an
indication of intrusion, central server communicates with the AP and orders
traffic capture and logs the intrusion. The capture will be analyzed later by the
fu
ll r
igh
ts.
WLAN administrator.
Keeping all this in mind, I did testing on 6 zones (18 APʼs) with 298 clients.
Each zone has around 30-70 clients. Clients were connecting randomly. Some APʼs
might have ACLʼs for authentication, but in my testing APʼs donʼt keep ACL lists.
ins
Clients use username and password for authentication. Each CPE sends its
eta
username and password to the AP after the request from AP. EAP is implemented
rr
where supplicant is CPE, authenticator is AP and authentication server is Radius
ut
ho
server. This is not a part of the setup and for some testing EAP is disabled. To be
07
,A
able solve this problem a small tool from Aegis can be installed to an XP Windows
20
box. Aegis is a tool which can help a client to generate EAP packets. It works fine,
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
but as soon as I installed Zyxel USB adapter, there were 3 wireless cards working in
te
Windows which led to numerous problems. Eventually I decided to take EAP out;
sti
tu
where I have to use my laptop as a rogue AP whereas in the other cases EAP was
NS
In
in place.
SA
Each AP has a built-in memory card on it. For traffic capture, tcpdump needs
to be installed in each AP. APʼs, which I used for testing, don't allow me to use their
©
wireless cards as both client and AP. Prism2.5 wireless cards have this problem. A
wireless card can be set either in a monitor mode or master mode. It canʼt work
both in monitor and master mode. Atheros chipset overcomes this problem by
supplying Virtual AP functionality.
Ibrahim Halil Saruhan
© SANS Institute 2007,
12
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
Basically, to be sure that the set-up works fine, we have to check each entity
in the system one-by-one. Central server should have the PHP interpreter installed
and also PHP code for the logic, database connectivity, and XML parsing and
fu
ll r
igh
ts.
sending notifications. SSH2 is needed for secure communication between APʼs and
the central server. Cron jobs should be running and have to make sure that
everything is running in their environment by using “env”. Also command line
interface “php.ini” should be modified according to the SSH2 package. APʼs should
ins
have site survey Perl script running and CPEʼs should also have their web server
Monitoring
ho
5.
rr
eta
running.
07
,A
ut
To be able to monitor the scenarios I developed a simple small GUI. Its code
is in the appendix.
©
SA
NS
In
sti
tu
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Ibrahim Halil Saruhan
© SANS Institute 2007,
13
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
The best monitoring environment is one which has sensors in place to detect
unauthorized activity; and services that can notify authorities via an alarm, analyze
data, and provide reports. (Lane, 2005)
fu
ll r
igh
ts.
I used site survey both from client view and AP view, Arp cache table and
CPE log table to do monitoring and collecting data. I leveraged iwlist while doing
site survey from AP and I set up a system by using PHP+SSH2+Perl and developed
a few programs to do site survey on the client. Clients canʼt run iwlist or any other
ins
wireless commands. I developed a few cron jobs which run periodically and in case
eta
of intrusion, APʼs are triggered to start logging traffic. In this section, AP and client
ho
rr
site survey methods are presented along with some analysis based on the data.
07
,A
ut
CPE Si te Surve y
20
I leveraged a Perl script, which basically creates an http request and sends it
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
to the CPE. CPE runs a small web server on it. It performs a site survey requested
tu
te
by the AP and responds with an http response packet. The same Perl script parses
sti
the response packet, gets the data out of it, creates an XML file with the site
NS
In
survey data, and sends it to the original requester which is central server.
SA
Here is a screen shot done on AP while the program traverses all the online
clients one by one and runs site survey on them. This data is collected and kept in
©
the database.
Ibrahim Halil Saruhan
© SANS Institute 2007,
14
As part of the Information Security Reading Room
Author retains full rights.
ut
ho
rr
eta
ins
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
07
,A
To be able to make it work continuously I created cron jobs like this:
4 ,23,41 =* AF19
* * * FA27
cpesitesurvey
env -i DE3D
PATH=/usr/bin/
php5A169
/var/www/secure.
Key fingerprint
2F94 998D FDB5
F8B5 06E4
4E46
In
sti
tu
te
20
GiacProject.net/Wids/CPESitesurvey8406502.php > /dev/null
6 ,25,44 * * * *
cpesitesurvey env -i PATH=/usr/bin/ php5 /var/www/secure.
GiacProject.net/Wids/CPESitesurvey8402001.php > /dev/null
8 ,27,46 * * * *
cpesitesurvey env -i PATH=/usr/bin/ php5
/var/www/secure.GiacProject.net/Wids/CPESitesurvey8409503.php > /dev/null
NS
Here is the table structure for the site survey data. For 298 clients running
SA
site survey 3 times in an hour, the number of records in the table was more than a
©
million after one week. Each site survey process runs from 12-18 seconds. In 20
minutes it can run up to 80 clients. In the testing system, none of the APʼs has that
many connected clients at any time.
CREATE TABLE `cpesitesurveys` (
Ibrahim Halil Saruhan
© SANS Institute 2007,
15
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
`id` INT NOT NULL AUTO_INCREMENT PRIMARY KEY,
`zoneid` int(7) unsigned zerofill NOT NULL default '0000000',
`customerid` INT(10) NOT NULL,
`site_ssid` varchar(50) default NULL,
`mac` VARCHAR(50) NOT NULL,
`channel` INT(10) NOT NULL,
`LogDate` datetime NOT NULL default '0000-00-00 00:00:00',
`signal` INT(10) NOT NULL
) TYPE = MYISAM;
eta
ins
fu
ll r
igh
ts.
Here is an excerpt from the CPE site survey table:
ho
rr
On the other hand, I can't run iwlist on clients but in AP's iwlist works fine.
07
,A
ut
CAU Site Surve y
CPE site survey is a few programs running together to collect data from
20
Key
fingerprint
= AF19on
FA27
998D
FDB5
DE3D F8B5
06E4 A169
4E46
clients.
CAU
site survey
the2F94
other
hand
is simple.
Basically,
I used
iwlist on the
tu
te
APʼs and recorded the responses to the database.
sti
Here is a screen shot done on AP. This data is collected and kept in the
©
SA
NS
In
database.
Ibrahim Halil Saruhan
© SANS Institute 2007,
16
As part of the Information Security Reading Room
Author retains full rights.
ins
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
eta
A beacon is a small broadcast data packet that reports the characteristics
rr
of the wireless network, with information such as supported data rate (max data
ut
07
,A
network name), etc. (Wong, 2004)
ho
rate), capabilities (encryption on or off), Access Point MAC address, SSID (wireless
When you check the response from iwlist, you can see the capability
20
Key fingerprint
= AF19
FA27
2F94For
998D
FDB5 DE3D
F8B5 06E4
A169 in
4E46
information
and some
more
data.
example,
capability
0X431
the second line
te
can be analyzed and it will show whether the AP supports WEP or not and also any
sti
tu
other features. All the capability bits can be checked by parsing this 0X431 value.
NS
In
Here is the table structure for the site survey data collected at the AP:
©
SA
CREATE TABLE `causitesurveys` (
`id` INT NOT NULL AUTO_INCREMENT PRIMARY KEY,
`channel` INT(10) NOT NULL,
`snr` INT(10) NOT NULL,
`capab` varchar(10) default NULL,
`rate` INT(10) NOT NULL,
`mac` VARCHAR(20) NOT NULL,
`suprates` VARCHAR(80) NOT NULL,
`zoneid` int(7) unsigned zerofill NOT NULL default '0000000',
`ssid` varchar(25) default NULL,
`LogDate` datetime NOT NULL default '0000-00-00 00:00:00'
) TYPE = MYISAM;
Here is an excerpt from the CAU site survey table.
Ibrahim Halil Saruhan
© SANS Institute 2007,
17
As part of the Information Security Reading Room
Author retains full rights.
rr
eta
ins
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
ho
Above, I tried to see whether someone is trying to spoof MAC address of
ut
one of the APʼs. The results were negative, as you canʼt see multiple SSIDʼs on the
07
,A
list.
20
Key
fingerprintto= the
AF19results
FA27 2F94
FDB5have
DE3DSNR
F8B5from
06E4 87
A169
According
our 998D
test APʼs
to4E46
53. After 53
te
tests shows other APʼs around. For this sample set we can say that all the 15 APʼs
sti
tu
tested are not affected by noise nor is there any MAC Spoofing.
In
From both the CPE and CAU site survey tables we can also observe that
NS
most of the users leave default names “Linksys, NETGEAR and belkin54g” on the
©
SA
wireless routers and most of these APʼs also donʼt support WEP.
Ibrahim Halil Saruhan
© SANS Institute 2007,
18
As part of the Information Security Reading Room
Author retains full rights.
6.
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
Detection
ins
This section starts by explaining 2 attacks namely Man-in-the-middle and Evil
eta
Twin attacks. After that I will give examples of MAC address list, spoofing, RF
ho
rr
jamming, de-authentication and authentication detection.
ut
My intention to do site survey in the first place was to be able to detect
07
,A
existence of an Evil Twin attack. I wasnʼt able to detect one actually. I wrote a
20
script where I can do it through adding data to the database, however I wasnʼt
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
satisfied. Therefore I bought a hardware USB adapter which can work as AP, went
tu
te
to the field and tried to create a rogue by myself and see whether the system
sti
would catch me. The experiment went smooth. My novice USB adapter turned to
In
be a good one and was detected by many clients. The program on the cron job
NS
notified the AP my existence and the AP automatically started capturing wireless
©
SA
traffic and logged it automatically to the central server.
Man-in-the-middle and Evil Twin attacks are pretty much similar attacks. Both
of them use a rogue AP and intend to create denial of service attack to the
legitimate AP. Evil Twin attack creates an evil twin by spoofing the MAC address
of the legitimate AP and also tries to forward the client to fakes login pages and
Ibrahim Halil Saruhan
© SANS Institute 2007,
19
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
trying to gather login information, whereas Man-in-the-middle attack tries to hide
its existence to work at least 5 channels away where the legitimate AP is
functioning and tries to listen the traffic.
fu
ll r
igh
ts.
I will start with what the others say about the attacks, showing the
experiment and itʼs results and end up with some analysis on what happened by
checking the wireless traffic.
ins
The strategy is simple. A cron job checks the CPE site survey table. In case it
eta
detects a rogue AP from this data, it triggers the AP to log the traffic. While
rr
capturing the traffic AP moves from master mode to monitor mode and then
ut
ho
back. The source code for all these steps is in the appendix.
07
,A
Man in the Middle At ta ck
20
Key
fingerprint can
= AF19
FA27 2F94 998D
FDB5 DE3D
F8B5 06E4 A169 4E46
An attacker
successfully
implement
a Man-in-the-middle
attack by first
te
configuring a rogue AP to imitate a legitimate AP, then coerce wireless clients to
sti
tu
connect to the rogue AP by performing a denial of service attack against the
In
legitimate AP or by providing a stronger signal than the targeted AP. Wireless
NS
clients will normally associate to the AP with the strongest signal or lowest signal
SA
to noise ratio (SNR). To make the intercepted connection appear seamless to
©
victims, the rogue AP could then bridge connections to another network
connection. If successfully executed an attacker will have complete control of the
wireless clientʼs network connection and may perform any inline attack they wish
(Deckerd, 2006)
AP impersonation attacks can be done for several purposes, including as a
Ibrahim Halil Saruhan
© SANS Institute 2007,
20
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
Man-in-the-middle attack, as a rogue AP attempting to bypass detection, and as a
possible honeypot attack. In such an attack, the attacker sets up an AP that
assumes the BSSID and ESSID of a valid AP. A successful Man-in-the-middle attack
fu
ll r
igh
ts.
will insert an attacker into the data path between the client and the AP. In such a
position, the attacker can delete, add, or modify data, provided he has access to
the encryption keys. Such an attack also enables other attacks that can learn a
userʼs authentication credentials. (Aruba Networks Technical Brief, 2007)
ins
After going to the field and running my USB wireless adapter as a rogue AP,
eta
I checked the existence of the rogue APʼs from the Monitor Program that I
rr
mentioned earlier. The program showed 2 cases, the second one being what I was
ut
ho
looking for. The first one actually shows that the MAC Address of the AP is
07
,A
changed because it is replaced. 00:02:6F series are 802.11b AP series and
©
SA
NS
In
sti
tu
te
20
00:15:6D series are the 802.11g series. Basically, the AP changed in that zone. This
Key also
fingerprint
= AF19
998D
FDB5 DE3D
06E4 list
A169
program
shoved
me FA27
that I2F94
need
to update
the F8B5
AP MAC
in 4E46
the system.
Right after the detection which we can see from the third and fourth lines,
Ibrahim Halil Saruhan
© SANS Institute 2007,
21
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
we see that it logged the event. The fourth line in the log below is actually the
incident traffic. I stopped the logging at 20.41 and at the time I started rogue AP
it logged it. I allowed it to log after the incident periodically for a while and moved
rr
eta
ins
fu
ll r
igh
ts.
it back to steady state again.
ho
When we check the database we see the multiple SSIDʼs in the table, which
07
,A
ut
triggered the whole process. The second line below is done manually by me to test
the cases. The third line actually is from the USB Adapter AP and the fourth line is
20
from the
theFDB5
legitimate
AP was
on channel 11 I
Key legitimate
fingerprint = AP.
AF19Even
FA27though
2F94 998D
DE3D F8B5
06E4working
A169 4E46
te
set up the rogue AP to work on channel 1. A Man-in-the-middle attack works when
SA
NS
In
sti
tu
the rogue AP is at least 5 channels away from the legitimate AP.
©
Here are the clients who noticed the rogue AP at the time of the incident.
The rogue AP worked for 23 minutes and 3 different clients noticed the existence
of the rogue AP. I didnʼt even boost the signal by using an external antenna.
Ibrahim Halil Saruhan
© SANS Institute 2007,
22
As part of the Information Security Reading Room
Author retains full rights.
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
After the incident I transferred the capture file “DUMP_8409503_
2_2007-07-22-21-32-15” to my laptop and opened it with Wireshark. I applied the
ins
filter “ wlan.sa == Ro gue AP MAC” or “ wlan.da = = R ogue AP MAC”
07
,A
ut
ho
rr
eta
Here is the screen shot:
©
SA
NS
In
sti
tu
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
We can see that rogue 00:13:49:AF:CB:53 Zyxel AP is sending beacons to
the network. Only one client, 00:02:6f:07:cf:f2, started to communicate with the
Ibrahim Halil Saruhan
© SANS Institute 2007,
23
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
rogue AP. In this traffic we can also notice that all the frames that the rogue AP is
sending are beacon frames. Beacon Frames can be transmitted by the AP for
polling purposes. The beacon frame sent by the AP contains control information
fu
ll r
igh
ts.
and can be used by mobile stations to locate an AP if it is on active scanning
mode. Briefly, rogue AP wants to let everyone know that it exists.
After identifying the only client connecting to the rogue AP I checked the
traffic originated from this client by using this filter “ wlan.sa ==
07
,A
ut
ho
rr
eta
ins
00 :0 2:6F:0 7:CF:F2 ”
©
SA
NS
In
sti
tu
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
After that I checked the subtype of the traffic for identification. After
adding “ wlan.fc.subt y pe! =4” filter all the probe request traffic is gone and only
re-association packets left.
Ibrahim Halil Saruhan
© SANS Institute 2007,
24
As part of the Information Security Reading Room
Author retains full rights.
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
ins
On the other hand, to be able to show the communication between the rogue
rr
eta
AP and legitimate client I used the following filter…
ho
“((wlan.sa == 00:13 :49:AF:C B :53) & & (wlan.da = =
ut
00 :0 2: 6f: 07 :cf :f 2)) | | ((wlan.da = = 00 :1 3:49 :AF:C B :5 3) & & (wlan.sa ==
07
,A
00 :0 2:6f:07 :cf :f 2))”
©
SA
NS
In
sti
tu
te
20
Key
fingerprint
= AF19shot
FA27of
2F94
FDB5 DE3D F8B5 06E4 A169 4E46
Here
is the screen
the998D
result:
Ibrahim Halil Saruhan
© SANS Institute 2007,
25
As part of the Information Security Reading Room
Author retains full rights.
ho
rr
eta
ins
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
ut
Here is a graphical representation of the connection establishment between
07
,A
an AP and a station.
©
SA
NS
In
sti
tu
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Th i s im a g e i s r et rie v e d f rom h tt p ://ww w . wa tc h g uard .c o m/arch ive/images/AP ha n d shake. jp g
From all the above I came to the following conclusion. The client is
continuously sending probe request packets to the rogue AP on a certain channel
Ibrahim Halil Saruhan
© SANS Institute 2007,
26
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
but the rogue AP is ignoring those requests. On the other hand the client went
through all the probe request/response, authentication, and association phases
with the legitimate AP, de-associated and now wants to re-associate again to start
fu
ll r
igh
ts.
transfer traffic. At this point the legitimate AP is down and the rogue AP doesnʼt
fulfill its request. AP in this case somehow locks the client from connecting to the
internet.
My intention was to implement Evil Twin attack but eventually I noticed that I
ins
essentially implemented Man-in-the-middle attack. The program detected Rogue
eta
AP right at the moment. The wireless traffic shows the existence of the rogue AP,
rr
struggling client sending probe requests and also trying to re-associate. Meanwhile,
ut
ho
the rogue AP is sending beacons. Also site survey shows that 3 clients see the
07
,A
rogue AP too. In case I had set up the rogue AP with the same channel of the
legitimate AP, I would have seen successful connections.
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
tu
te
E vi l Twin Attac k
sti
Here are the SSIDʼs and the total number of occurrences in the CPE site
©
SA
NS
In
survey table.
Both queries below returned nothing. There was no indication of Evil Twin
Ibrahim Halil Saruhan
© SANS Institute 2007,
27
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
attack. What I was trying to do was finding multiple logs for the same exact time
from the same client with the same MAC address. It returned nothing, which means
ins
fu
ll r
igh
ts.
there werenʼt any real Evil Twin attacks occurring in those 18 APʼs I was testing.
eta
Another method for detecting Evil Twin attack would be to use kismet to
rr
find evil twins. PRISM2 header contains Received Signal Strength Indication (RSSI)
ho
for every packet. RSSI is a measurement of the received radio signal strength. It is
ut
possible to log the traffic and use Wireshark to see the signal on the packets. We
07
,A
need per-packet signal strength for that to work because it watches all the beacon
te
tu
MAC Address List Match
20
and probe frames and keeps track of their signal strength.
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
sti
In order to detect rogue access points, the IDS utilize a list of authorized
In
access points then alerts when a detected AP does not match the list. (Vladimirov,
SA
NS
Gavrilenko, and Mikhailovsky, 2004)
©
I kept the lists in the database at the Central Server.
Client MAC Address Match
I used Arp on APʼs and a MAC table for clients in the central server. I had a
cron job running every 5 minutes which sends an email notification if a MAC
Ibrahim Halil Saruhan
© SANS Institute 2007,
28
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
address in the Arp table on the AP doesnʼt match with the database. I created an
email account as GiacMonitor@gmail.com and turned of the authentication for a
07
,A
ut
ho
rr
eta
ins
fu
ll r
igh
ts.
while and this was the result.
As we can see from the e-mail someone is connected to the AP and we can
te
20
Keywith
fingerprint
= AF19
2F94 998D
FDB5
DE3D F8B5 06E4 A169 4E46
also tell
proximity
ofFA27
5 minutes
when
it happened.
sti
tu
AP MAC Address Match
In
Another attribute that can be implemented on the wireless IDS is to create a
NS
listing or be pre-configured with all known and authorized WAPs, so that whenever
SA
an unidentified or rogue WAP is found in the network, the wireless IDS can quickly
©
detect and alert. (Poblete, 2005)
For AP MAC Address Match, I used MAC addresses of the APʼs from CPE
site survey table and an AP MAC List table in the database at the central server. As
long as the list is green there is no indication of spoofing.
Ibrahim Halil Saruhan
© SANS Institute 2007,
29
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
07
,A
ut
ho
rr
eta
ins
fu
ll r
igh
ts.
Here is a screen shot of the MAC address checking for the AP:
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
te
MAC Address Sp oof ing
sti
tu
Impersonation attacks in a wireless network typically involve an attacker
In
taking on the address of a valid client or AP and trying to obtain access or
NS
services typically reserved for those valid clients or APs. (Aruba Networks
SA
Technical Brief, 2007)
©
MAC address Spoofing is very easy to do. The following command will let
someone to change the MAC address of their network card.
ifconfig eth1 hw ether AA:BB:CC:DD:00:11
Ibrahim Halil Saruhan
© SANS Institute 2007,
30
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
Where eth1 is the network interface and AA:BB:CC:DD:00:11 is the MAC
address.
The test program checks multiple MAC addresses in the Arp table on the AP.
fu
ll r
igh
ts.
I created a cron job and checked the logs but didnʼt see a single MAC address
spoofing case. I believe people do not bother to sniff the wireless traffic, grab the
MAC address of a connected client and try to connect simultaneously or wait until
that client finishes its connection and connect afterwards. The only time a rogue
ins
client can not be detected is when it spoofs a legitimate MAC address of a wireless
eta
network and when the actual client is not connecting. In this case there is no
rr
authentication like 802.1x. As long as the list is green there is no indication of
07
,A
ut
ho
spoofing.
In
sti
tu
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
SA
NS
RF Jamming
©
RF Jamming is counted as DOS attacks by many, however I think DOS
attacks are intentional, but RF Jamming can be unintentional by nature. Therefore I
didnʼt show RF Jamming as a part of DOS attacks.
Baby monitors and other devices that operate on the 2.4 GHz band can
disrupt a wireless network using this frequency. These denials of service can
Ibrahim Halil Saruhan
© SANS Institute 2007,
31
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
originate from outside the work area serviced by the access point, or can
inadvertently arrive from other 802.11b devices installed in other work areas that
degrade the overall signal. (Internet Security Systems White Paper, 2007)
fu
ll r
igh
ts.
Disruptions can be caused by noise from microwaves, cordless phones, or
other appliances that operate on the 2.4 GHz radio frequency on which 802.11b
wireless LANs also operate. Disruptions can also be caused by hackers using
ins
access points to send dissociate commands. (Lane, 2005)
eta
RF technology is very dynamic in nature as it changes in response to the
rr
real-world conditions. Noise, interference, and increased traffic load, signal
ho
attenuation ‒ these are some of the factors that can cause RF topology to change
ut
from one moment to the next. For most network managers RF technology has
07
,A
remained something that is always in the air. (Anand, 2004)
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
RF jamming is used to take down an entire wireless LAN by overwhelming the
tu
te
radio environment with high-power noise. Channel or power adjustments to
In
sti
overcome the interference can be done (Aruba Networks Technical Brief, 2007)
NS
I developed 2 small programs to check noise both on the CPE level and CAU
SA
level. I leveraged another Perl script which also is doing site survey and logs the
data in a CPE log table. The main difference between this table and the CPE site
©
survey table is that this can also check the noise around it. Here are two screen
shots from these logs showing noisy APʼs and clients. All the screen shots shown in
this paper have small programs behind them and the source code is attached in
the appendix.
Ibrahim Halil Saruhan
© SANS Institute 2007,
32
As part of the Information Security Reading Room
Author retains full rights.
07
,A
ut
ho
rr
eta
ins
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
©
SA
NS
In
sti
tu
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
DOS At t acks
DOS attacks are designed to prevent or inhibit legitimate users from
accessing the network. This includes blocking network access completely,
degrading network service, and increasing processing load on clients and network
Ibrahim Halil Saruhan
© SANS Institute 2007,
33
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
equipment. (Aruba Networks Technical Brief, 2007)
Denial of service attacks are also easily applied to wireless networks, where
legitimate traffic can not reach clients or the access point because illegitimate
fu
ll r
igh
ts.
traffic overwhelms the frequencies. An attacker with the proper equipment and
tools can easily flood the 2.4 GHz frequency, corrupting the signal until the
wireless network ceases to function. (Internet Security Systems White Paper, 2007)
ins
Saturating an AP or client with requests is commonly used to deny service.
eta
To ensure WLAN high availability, flooding attacks need to be detected. These
rr
attacks include associate, re-associate, and disassociate requests, probe requests,
ho
and other reserved management frame flooding (Subtypes 6, 7, D, E, and F).
ut
(3Com Corporation White Paper, 2007) Some of the common DoS attacks are RF
07
,A
jamming attack, FataJack attack, Duration attack, Authentication storm, De-
te
20
authentication
storm,
Association
Disassociation
etc.4E46
(Anand, 2004)
Key fingerprint
= AF19
FA27 2F94storm,
998D FDB5
DE3D F8B5 storm
06E4 A169
tu
Duration Attack
In
sti
When a client sends frames with prolonged duration, other clients in the
NS
network have to wait till the specified duration to use the RF medium. If the client
SA
continuously sends frames with such high duration, then it can prevent other clients
from using RF medium and remain unassociated forever. WLAN devices perform
©
virtual carrier sensing prior to using the RF medium. Carrier sense minimizes the
likelihood of two devices transmitting simultaneously. Wireless nodes reserve the
right to use the radio channel for the duration specified in the frame. (Anand,
2004)
Ibrahim Halil Saruhan
© SANS Institute 2007,
34
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
The duration value in the frame indicates the duration in milliseconds for
which the channel is reserved. The Network Allocation Vector (NAV) stores this
duration information and is traced for every node. The basic rule is that any node
fu
ll r
igh
ts.
can transmit only if the NAV reaches zero or in other words no one has reserved
the channel at that time. Attackers take advantage of the NAV. An attacker can
send frames with huge duration values. This would force other nodes in the range
ins
to wait till the value reaches zero. (Anand, 2004)
eta
Deauthentication
rr
Spoofed deauthenticate frames form the basis for most denial of service
ho
attacks, as well as the basis for many other attacks such as Man-in-the-middle
ut
attack. A Linux driver called AirJack typically forms the basis for this type of
07
,A
attack, with tools such as WLAN-Jack and Fake-Jack actually carrying out the
20
attack. Broadcast deauthentication attack generates spoofed deauthenticate
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
frames with a broadcast destination address instead of disconnecting a single
tu
te
station, the intent is to disconnect all stations attached to a given AP. Typically, a
sti
Linux tool known as “Hunter-Killer” is used to generate this attack. FakeAP is a tool
In
originally created to thwart war drivers by flooding beacon frames containing
NS
hundreds of different addresses. This would appear to a war driver as though there
SA
were hundreds of different APs in the area, thus concealing the real AP. (Aruba
©
Networks Technical Brief, 2007)
Deauthentication attacks are one of the most common DOS attacks. This is
usually the initial step an attacker takes to create DOS attack against an AP with
the intention to make it not respond and force the clients to use the rogue AP. I
Ibrahim Halil Saruhan
© SANS Institute 2007,
35
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
implemented it on one of the test APʼs by using aireplay-ng.
Here is the command.
fu
ll r
igh
ts.
aireplay-ng -0 1 -a 00:14:6C:7E:40:80 -c 00:0F:B5:34:30:30 ath0
Where:
-0 means deauthentication
•
1 is the number of deauths to send (you can send multiple if you wish); 0
means send them continuously
•
-a 00:14:6C:7E:40:80 is the MAC address of the access point
•
-c 00:0F:B5:34:30:30 is the MAC address of the client to deauthenticate; if
this is omitted then all clients are deauthenticated
•
ath0 is the interface name
07
,A
ut
ho
rr
eta
ins
•
(Aircrack-ng Contributors, 2007)
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
te
I used a live BackTrack2 CD to implement the attack. I believe it is the
tu
simplest way of running the attack. I had both Atheros (installed by default) and
In
sti
external Prism card in my laptop. After booting from a live BackTrack2 CD, I set
NS
up my Prism card to monitor and my Atheros card to create the attack.
SA
While in RFMON mode, wireless clients are unable to transmit any frames;
©
their cards are only able to receive, and therefore capture traffic. This limits the
client to reporting only current or recorded network traffic. For instance, a client
using passive monitoring would be able to report on the MAC addresses and
number of associations to a discovered AP, but would be unable to probe the
discovered AP for SNMP MIB information. (Wright, 2002)
Ibrahim Halil Saruhan
© SANS Institute 2007,
36
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
I put my Prism card to monitor mode like this:
iwconfig eth1 mode monitor
tcpdump
fu
ll r
igh
ts.
And after that I started the tcpdump packet capture for analysis:
-i eth1 –w DeAuthtest
Meanwhile I destroyed my Atheros card interface first, then created a new
virtual interface assigning it to wifi0 and putting it in the monitor mode like this
ins
bt ~ # wlanconfig ath0 destroy
bt ~ # wlanconfig ath0 create wlandev wifi0 wlanmode monitor
eta
I checked whether the card is up and made sure that it is up and in monitor
ho
rr
mode.
ut
bt ~ # ifconfig ath0 up
bt ~ # iwconfig
A169 4E46
tu
te
20
07
,A
ath0
IEEE 802.11g ESSID:"" Nickname:""
Mode:Monitor Frequency:2.462 GHz Access Point: 00:19:7D:4A:97:96
Bit Rate:0 kb/s
Tx-Power:31 dBm
Sensitivity=0/3
Retry:off
RTS thr:off
Fragment thr:off
Encryption key:off
Key
fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4
Power Management:off
Link Quality=0/94 Signal level=-95 dBm Noise level=-95 dBm
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc:0
Missed beacon:0
sti
I moved to the folder where aireplay-ng is and started the attack. -0 means
In
that it is a deauthentication attack and 0 means that it will be continuously. On the
NS
other hand, ‒a means the APʼs MAC address and ‒c means the clientʼs MAC
SA
address and ends with the interface that the attack would generate from. To be
©
able to start the attack the interface should be up and in monitor mode. Omitting ‒
c will make the attack broadcast and it effects all the connected clients at the AP.
bt ~ # cd /pentest/wireless/aircrack-ng/
bt aircrack-ng # aireplay-ng -0 0 -a 00:15:6D:63:11:DC -c 00:19:7E:0B:71:8F ath0
19:04:29
19:04:30
Sending DeAuth to station
Sending DeAuth to station
-- STMAC: [00:19:7E:0B:71:8F]
-- STMAC: [00:19:7E:0B:71:8F]
Ibrahim Halil Saruhan
© SANS Institute 2007,
37
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
19:04:31
19:04:32
19:04:34
19:04:35
19:04:36
19:04:37
Sending
Sending
Sending
Sending
Sending
Sending
DeAuth
DeAuth
DeAuth
DeAuth
DeAuth
DeAuth
to
to
to
to
to
to
station
station
station
station
station
station
-------
STMAC:
STMAC:
STMAC:
STMAC:
STMAC:
STMAC:
[00:19:7E:0B:71:8F]
[00:19:7E:0B:71:8F]
[00:19:7E:0B:71:8F]
[00:19:7E:0B:71:8F]
[00:19:7E:0B:71:8F]
[00:19:7E:0B:71:8F]
Here is Wireshark capture that the attack is actually working against the
07
,A
ut
ho
rr
eta
ins
fu
ll r
igh
ts.
client 00:19:7E:0B:71:8F
NS
In
sti
tu
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
SA
801 .1x EAP Authenticat i on
©
A skilled attacker captures wireless traffic using techniques similar to those
employed on wired networks. Many of these tools capture the first part of the
connection session, where the data would typically include the username and
password. An intruder can then masquerade as a legitimate user by using this
Ibrahim Halil Saruhan
© SANS Institute 2007,
38
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
captured information to hijack the user session and issue unauthorized commands.
(Internet Security Systems White Paper, 2007)
I logged in to one of the test APʼs and forced it to reboot. On another AP I
fu
ll r
igh
ts.
captured the traffic. As we can see after the initial 3 steps (Probe Request and
Response, Authentication, Association) EAP Authentication phase starts. AP starts
the authentication and after it gets the MD-5 Challenge it checks it and replies with
07
,A
ut
ho
rr
eta
ins
a success message.
Prevention
©
7.
SA
NS
In
sti
tu
te
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Once a rogue AP is discovered the next immediate step is to block the AP
from the network so that the authorized clients donʼt associate with it. There are
two ways of blocking the rogue APs. 1. Tit for Tat: Launch a Denial-of-service
(DOS) attack on the rogue AP and make it deny wireless service to any new client.
Ibrahim Halil Saruhan
© SANS Institute 2007,
39
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
2. Pull it out of the network: Either the WLAN administrator can manually locate
the AP and pull it physically off the LAN OR block the switch port to which the AP
is connected. (Anand, 2004)
fu
ll r
igh
ts.
In our testing case choice number 1 is reasonable. By using tools like void11,
hunter-killer or aireplay-ng we can create DOS attack on the rogue AP so that we
can make it sure that it denies wireless service to any new client.
ins
The access control list is the simplest security measure we can find in a
eta
wireless network. The protection offered by this mechanism mainly consists of
rr
filtering out unknown users and requires a list of authorized clientʼs MAC addresses
ho
to be loaded in the Access Point. Only those registered MAC addresses will be able
ut
to communicate with the Access Point, and will drop any communication that come
07
,A
from others not registered MAC addresses. (Wong, 2004)
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Once a rogue client is detected, WLAN administrator should shut down the
tu
te
client from the network. The most common method of keeping rogue clients away
sti
is by configuring their MAC address in the Access Pointʼs Access Control List
In
(ACL). ACL determines whether to deny or allow a client to connect to the AP.
NS
WLAN administrators can specify the rogue clientʼs MAC address in the ACL of all
©
(Anand, 2004)
SA
authorized Access Points to keep the rogue client off the network for ever.
Keeping ACLʼs in an AP is a way of keeping rogue clients away. This project
on the other hand checks Arp cache table, getʼs the IP-MAC pairs and compare
them with the MAC addresses from the database at the Central Server. So the
difference is between keeping the list in AP or a Central Server. I think both ways
Ibrahim Halil Saruhan
© SANS Institute 2007,
40
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
are pretty much protective.
802.11 frames are trivial to forge. WPA RADIUS looks the most secure
solution with PEAP as the authentication method or EAP-TLS, which are
fu
ll r
igh
ts.
equivalently secure.
Infrastructure mode is the most common operation mode in which we could
find wireless networks. In this operation mode, each wireless client connects
eta
rr
between others wireless clients. (Wong, 2004)
ins
directly to a central device called Access Point; there is no direct connection
ho
Two wireless clients can talk directly to each other, bypassing the access
ut
point. Users therefore need to defend clients not just against an external threat but
07
,A
also against each other. (Internet Security Systems White Paper, 2007)
20
Key
fingerprintclients
= AF19to
FA27
998D other
FDB5 DE3D
F8B5
06E4 A169
4E46 Client
Preventing
talk2F94
to each
is called
Client
Isolation.
te
isolation should be activated on the AP. It prevents wireless clients from talking to
sti
tu
one another. The typical way for someone to attack one of the clients is to inject
In
packets into the AP that the AP dutifully repeats to the client and they should be
NS
prevented from doing so because they have no business talking to each other
SA
directly. This is also called as "AP isolation”. In our system it is implemented by
using hostapd. This way an attacker can now only attack the AP. To attack the
©
client directly they'd have to spoof frames from the AP and be close enough to
the client that the client can hear them which significantly reduces the attacker's
range and potential targets
Clever deployers of rogue access points have been known to purchase
Ibrahim Halil Saruhan
© SANS Institute 2007,
41
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
unauthorized 802.11a devices on the theory that the existing network is not
capable of detecting them. (Gast, 2005)
The current set-up in this paper recognizes only 802.11b and 802.11g APʼs
fu
ll r
igh
ts.
because it works on the open 2.4 GHz frequency. 802.11g is an upgrade to
802.11b, and supports up to 54Mbps. 802.11 also supports 54Mbps, but on 5.0
GHz, which makes it impossible for the 802.11b CPEʼs to recognize it. There should
ins
be other measures to detect 802.11a APʼs.
eta
When a client exceeds a pre-specified rate of 802.11 associate, disassociate,
rr
or re-associate packets, it should automatically be put on a blacklist for a pre-
ho
specified amount of time to allow an administratorʼs assessment of the situation.
ut
Since spoofed de-authenticate frames are the basis for DOS and Man-in-the-middle
07
,A
attacks, detecting these frames provides a security alert. Detecting probe
20
responses
containing
a null
SSID
that
willFDB5
disable
a number
ofA169
popular
Key fingerprint
= AF19
FA27
2F94
998D
DE3D
F8B5 06E4
4E46 network
tu
te
interface cards can also deter attacks. (3Com Corporation White Paper, 2007)
sti
Most client intrusion attempts are handled by higher-layer security functions.
In
However, one serious lower-layer attack that exploits client weaknesses is the
NS
honeypot AP. A “honeypot” has a number of connotations in the security world.
SA
When discussing wireless LANs, one meaning is an attackerʼs AP that is set up in
©
close proximity to an enterprise, advertising the ESSID of the enterprise. The goal
of such an attack is to lure valid clients to associate to the honeypot AP. From
that point, a Man-in-the-middle attack can be mounted, or an attempt can be made
to learn the clientʼs authentication credentials. Most client devices have no way of
distinguishing between a valid AP and an invalid one ‒ the devices only look for a
Ibrahim Halil Saruhan
© SANS Institute 2007,
42
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
particular ESSID and will associate to the nearest AP advertising that ESSID.
(Aruba Networks Technical Brief, 2007)
In my testing system, each AP has a different ESSID based on their location.
fu
ll r
igh
ts.
Clients perform site surveys and collect data about APʼs around them. This data is
stored in the database. A rogue AP will be recognized by multiple clients and site
survey table will show whether it is an AP with a different MAC Address, possibly
indicating a Man-in-the-middle attack, or a rogue AP seen by many clients at the
ins
same time and same MAC address, which most probably indicates the existence of
rr
eta
the Evil Twin attack.
ho
Rogue detection and blocking is a continuous process involving at least
07
,A
ut
three components:
te
20
• A dedicated piece of hardware probe/sensor to monitor the air and
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
identify network behavior
sti
tu
• A central IDS engine that gathers inputs from many such probes/sensors
In
and helps in pinpointing a device as rogue.
NS
• A network management software that can talk to the wired network,
SA
identify the switch port to which the rogue AP is connected and shutdown the
©
port.
(Anand, 2004)
Ibrahim Halil Saruhan
© SANS Institute 2007,
43
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
8.
C onclu sion
Site survey on the client side is a very effective method. By using this data
Iʼm able to implement rogue device detection without installing dedicated devices
fu
ll r
igh
ts.
on the network.
All the other techniques like MAC list checking, MAC spoofing checking, and
ins
RF jamming checking are additional tools to detect rogues on the network.
eta
I believe this set-up allows me to watch as many APʼs as I want for rogues. It
rr
is scalable. This paper approached the problem of rogue detection and prevention
ho
from a different perspective.
07
,A
ut
Actually, I checked the database for the weeks of logs and didnʼt see a single
rogue AP which has a spoofed MAC address. I also checked the CPE site survey
20
fingerprint
= AF19
FA27
2F94
998D for
FDB5
DE3D
F8B5
A169
resultsKey
coming
from
almost
300
clients
more
than
2 06E4
weeks
and4E46
didnʼt find a
te
single case there either. I also checked for other well known default SSIDʼs like
sti
tu
Linksys and NETGEAR and I wasnʼt able to see a single case there either.
In
I think Rogue APʼs are mostly implemented in airports or Starbucks coffee
NS
shops. I was checking connectivity in the airport yesterday and wasnʼt able to
SA
connect any AP. I decided to change the mode of my wireless USB adapter from
©
station mode to AP mode and used the broadcast SSID t-mobile. Suddenly, I saw
connecting clients on the MAC list of my AP. Here is the screen shot of it:
Ibrahim Halil Saruhan
© SANS Institute 2007,
44
As part of the Information Security Reading Room
Author retains full rights.
eta
ins
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
rr
I did a few experiments on the field by using my USB wireless adapter and
ut
ho
created rogue APʼs and detected them. The solution is working as it is intended to
Key
= AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
F u tfingerprint
ure Work
20
9.
07
,A
work.
tu
te
Detection of the physical location of an attack is a critical aspect of a
sti
wireless IDS. For the most common standard 802.11, attacks are often carried out
In
from a close proximity to a WAP, and most likely are carried out in an extremely
SA
NS
short period to further avoid detection. (Poblete, 2005)
©
With the current set-up we can guess where the rogue device is, but a
sophisticated solution to the project which allows the system not only detect the
rogue, but also detect itʼs physical location would be a nice addition.
By using tools like void11 or hunter-killer we can create DOS attack on the
rogue AP. I believe it would be a nice addition to test each of these along with
Ibrahim Halil Saruhan
© SANS Institute 2007,
45
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
other tools. After that, we should check the site survey and logged traffic with the
proposed system. Some of these tools are card dependent. For example void11
works with Prism cards but not with Atheros cards.
fu
ll r
igh
ts.
The Snort-Wireless project is an attempt to make a scalable, and not to
mention free 802.11 intrusion detection system that can easily be integrated into
an IDS infrastructure. It is completely backwards compatible with Snort 2.0.x and
adds several additional features. Currently it allows for 802.11 specific detection
ins
rules through the new "WiFi" rule protocol, as well as rogue AP, AdHoc network,
rr
eta
and Netstumbler detection (Poblete, 2005)
ho
Snort-wireless is a wireless IDS designed to integrate into a Snort 2.x
ut
environment. Snort is the most widely deployed open source IDS, so a wireless
07
,A
plug-in makes sense for many enterprises. Snort-wireless allows for custom rules to
20
be created
based on
framing
from
a wireless
packet.
It also contains
Key fingerprint
= AF19
FA27information
2F94 998D FDB5
DE3D
F8B5 06E4
A169 4E46
te
rules to attempt to find rogue access points, wardrivers, and ad hoc networks.
sti
tu
(Potter, 2007)
In
The next step might be adding Snort-Wireless to the current project. Site
NS
survey, MAC address list control, spoofing control and all the other cases already
SA
logs the intrusions, but they can be more centralized, log the traffic and
©
automatically forward the traffic to snort-wireless.
Monitoring interface should hop between the 12 channels available to
wireless networks. Several wireless attacks work by utilizing a rogue AP on a
different channel. For instance monkey in the middle attacks utilize a rogue AP that
is at least 5 channels away from the target AP. Without channel hoping wireless IDS
Ibrahim Halil Saruhan
© SANS Institute 2007,
46
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
would be blind to attacks that function on other channels. (Deckerd, 2006)
This project doesnʼt do channel hopping while logging the traffic. However
site survey functionality logs the channel data. In case the system detects a rogue
fu
ll r
igh
ts.
AP, it will also know on which channel the rogue is working. It currently logs the
traffic on the channel of the legitimate AP; however it would be a nice addition to
log the traffic on all channels at the time of intrusion.
eta
ins
10. References
rr
1. Ellingson, Jorgen. (2001). Layers One &Two of 802.11 WLAN Security.
ho
Retrieved July 22,2007, from
07
,A
ut
http://www.giac.org/certified_professionals/practicals/GSEC/0996.php
20
2. Aircrack-ng Contributors. (07/22/2007). Aircrack-ng Deauthentication .
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Retrieved July 22,2007, from http://www.aircrack-
tu
te
ng.org/doku.php?id=deauthentication
In
sti
3. Gast, Matthew. 802.11Wireless Networks: The Definitive Guide. Sebastopol,
NS
CA: O'Reilly, 2005
SA
4. Vladimirov, Andrew A., Konstantin V. Gavrilenko, and Andrei A. Mikhailovsky.
©
Wi-Foo: The Secrets of Wireless Hacking. Boston: Addison-Wesley, (2004)
5. Lane, Heather D. (02/6/2005). Security Vulnerabilities and Wireless LAN
Technology. Retrieved July 22,2007, from
http://www.giac.org/certified_professionals/practicals/GSEC/4383.php
Ibrahim Halil Saruhan
© SANS Institute 2007,
47
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
6. Wireless LANs: Assuring Enterprise Security and Identity Awareness.
(07/22/2007). 3Com Corporation White Paper. Retrieved July 22,2007,
from
fu
ll r
igh
ts.
http://www.3com.com/corpinfo/en_US/technology/tech_paper.jsp?DOC_ID=
230943
7. Potter, Bruce. (07/22/2007). Wireless Intrusion Detection. Retrieved July
ins
22,2007, from http://www.itsec.gov.cn/webportal/download/88.pdf
eta
8. Poblete, Oliver. (01/24/2005). An Overview of the Wireless Intrusion
rr
Detection System. Retrieved July 22,2007, from
ut
ho
http://www.giac.org/certified_professionals/practicals/GSEC/4296.php
07
,A
9. Wireless LAN Security 802.11b and Corporate Networks. (07/22/2007).
tu
Anand, Dev. (2004). Effective WLAN Management With Distributed RF
sti
10.
te
20
Internet Security Systems White Paper. Retrieved July 22,2007, from
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
http://www.iss.net/documents/whitepapers/wireless_LAN_security.pdf
In
Sensors. An Adventnet Technical Whitepaper. Retrieved July 22,2007, from
NS
http://manageengine.adventnet.com/products/wifi-
Anand, Dev. (2004). Rogue Detection and Blocking. An Adventnet
©
11.
SA
manager/rfsensor_whitepaper.pdf
Technical Whitepaper. Retrieved July 22,2007, from
http://manageengine.adventnet.com/products/wifi-manager/roguedetection-and-blocking.pdf
Ibrahim Halil Saruhan
© SANS Institute 2007,
48
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
12.
Wright, Joshua. (11/08/2002). Layer 2 Analysis of WLAN Discovery
Applications for Intrusion Detection. Retrieved July 22,2007, from
http://www.rootsecure.net/content/downloads/pdf/wlan_ids.pdf
Deckerd, Gary. (11/23/2006). Wireless Attacks from an Intrusion
fu
ll r
igh
ts.
13.
Detection Perspective. Retrieved July 22,2007, from
http://www.giac.org/certified_professionals/practicals/GCIA/0878.php
Hutchison, Ken. (10/18/2004). Wireless IntrusionDetection Systems.
ins
14.
eta
Retrieved July 22,2007, from
Wong, Luis Carlos. (10/21/2004). An Overview of 802.11 Wireless
ut
15.
ho
rr
http://www.giac.org/certified_professionals/practicals/GSEC/4210.php
07
,A
Network Security Standards & Mechanisms. Retrieved July 22,2007, from
Gast, Matthew. 802.11Wireless Networks: The Definitive Guide.
te
16.
20
http://www.giac.org/certified_professionals/practicals/GSEC/4185.php
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Wireless Intrusion Protection (07/22/2007). Aruba Networks Technical
In
17.
sti
tu
Sebastopol, CA: O'Reilly, 2005
NS
Brief. Retrieved July 22,2007, from
18.
©
SA
http://www.arubanetworks.com/pdf/technology/tb_wip.pdf
Beacon Frame. (07/22/2007). In Wikipedia, The Free Encyclopedia.
Retrieved July 22, 2007, from http://en.wikipedia.org/wiki/Beacon_frame
19.
Phifer, Lisa. (07/22/2007). Anatomy of a Wireless "Evil Twin" Attack.
Retrieved July 22, 2007, from
Ibrahim Halil Saruhan
© SANS Institute 2007,
49
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
http://www.watchguard.com/infocenter/editorial/27061.asp
11. Appendix
fu
ll r
igh
ts.
S ource C ode
Client MAC Address List Check
function CPEMacAddressCheck($CpeZoneId){
eta
ins
$result
= mysql_fetch_assoc(mysql_query("SELECT NasIPAddress as CauIpAddress FROM
RadiusNas WHERE (zone = $CpeZoneId)"));
$CauIpAddress = $result[CauIpAddress];
$conn
= ssh2_connect($CauIpAddress,22);
ho
rr
if (ssh2_auth_password($conn,'UserName','Password'))
echo "<br><center>Connected to the CAU at $CpeZoneId.</center><br>";
else
die('<br><center>Authentication is Failed...</center><br>');
07
,A
ut
$arpcommand
= "
/usr/sbin/arp -e | grep '172.16.[4-6]' | sort | /usr/bin/awk
'/ether/ {print $3}'";
$stream
= ssh2_exec($conn, $arpcommand);
$Dump
= '';
print "<table align= 'center' border= '1px solid #BDBABD' border-collapse='collapse'
bgcolor='#F7F7F7'>";
20
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
SA
}
NS
In
sti
tu
te
while($line = fgets($stream)){
$line = trim(strtolower($line));
$result = mysql_query("SELECT count(*) as TOTAL FROM MacList WHERE CPE_MAC =
'$line'");
$note = mysql_fetch_assoc($result);
if ($note[TOTAL] > 0)
$line = '<font color="green">'
. $line . '</font>' ;
else
$line = '<font color="red">' . $line . '</font>' ;
$Dump = $Dump . '<tr><td>' . $line . '</td></tr>' ;
}
print "$Dump";
print "</table>";
©
Client MAC Address Spoofing Check
function CPESpoofingCheck ($CpeZoneId){
$result
= mysql_fetch_assoc(mysql_query("SELECT NasIPAddress as CauIpAddress FROM
RadiusNas WHERE (zone = $CpeZoneId)"));
$CauIpAddress = $result[CauIpAddress];
$conn
= ssh2_connect($CauIpAddress,22);
if (ssh2_auth_password($conn,'UserName','Password'))
Ibrahim Halil Saruhan
© SANS Institute 2007,
50
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
echo "<br><center>Connected to the CAU at $CpeZoneId.</center><br>";
else
die('<br><center>Authentication is Failed...</center><br>');
fu
ll r
igh
ts.
$arpcommand
= "
/usr/sbin/arp -e | grep '172.16.[4-6]' | /usr/bin/awk '/ether/
{print $3}' | sort";
$stream
= ssh2_exec($conn, $arpcommand);
$Dump
= '';
print "<table align= 'center' border= '1px solid #BDBABD' border-collapse='collapse'
bgcolor='#F7F7F7'>";
$previousline = 'none';
eta
ins
while($line = fgets($stream)){
$line = trim(strtolower($line));
if ($line == $previousline)
$line = '<font color="red">' . $line . '</font>' ;
else
$line = '<font color="green">'
. $line . '</font>' ;
$Dump
= $Dump . '<tr><td>' . $line . '</td></tr>' ;
$previousline = $line;
}
print "$Dump";
print "</table>";
rr
}
ut
ho
AP MAC Address List Check
$zoneid
$radio
$ReturnValue
07
,A
function CAUMacAddressCheck ($site_ssid, $mac){
= substr($site_ssid, 4, 7);
= substr($site_ssid, 12, 1);
= false;
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
1)
2)
3)
4)
$tempRadio
$tempRadio
$tempRadio
$tempRadio
=
=
=
=
'radio1_mac';
'radio2_mac';
'radio3_mac';
'radio4_mac';
20
==
==
==
==
te
($radio
($radio
($radio
($radio
tu
if
if
if
if
NS
In
sti
$SQLQuery
= "SELECT" . ' ' . $tempRadio . ' ' . "FROM serial WHERE ssid = '$zoneid'
ORDER BY install_date DESC";
$result
= mysql_query($SQLQuery);
$note
= mysql_fetch_assoc($result);
if (strtoupper($note[$tempRadio]) == strtoupper($mac))
$ReturnValue = true;
return ($ReturnValue);
}
SA
function CAUMacAddressControl(){
©
$result
= mysql_query("SELECT * FROM cpesitesurveys WHERE site_ssid like '%BBC%'
OR site_ssid like '%TBZ%' GROUP BY mac ORDER BY site_ssid");
print "<table align= 'center' border= '1px solid #BDBABD' border-collapse='collapse'
bgcolor='#F7F7F7'>";
print "<tr><td align='center' colspan=4><b>Mac Address Control</b></td></tr>";
while($note = mysql_fetch_assoc($result))
if (CAUMacAddressCheck ($note[site_ssid], $note[mac]))
print "<tr><td><b>SSID</b></td><td><font
color='green'><b>$note[site_ssid]</b></td><td><b>Mac
Address</b></td><td><font
color='green'><b>$note[mac]</b></font></td></tr>";
else
Ibrahim Halil Saruhan
© SANS Institute 2007,
51
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
print "<tr><td><b>SSID</b></td><td><font
color='red'><b>$note[site_ssid]</b></td><td><b>Mac
Address</b></td><td><font color='red'><b>$note[mac]</b></font></td></tr>";
print "</table>";
}
fu
ll r
igh
ts.
Main Monitor
include ("wids/EvilTwin.func.php");
include ("wids/CAUMacAddressControl.func.php");
include ("wids/CPEMacAddressControl.func.php");
include ("wids/RFJamming.func.php");
include ("wids/Traffic.func.php");
©
SA
NS
In
sti
tu
te
20
07
,A
ut
ho
rr
eta
ins
if(isset($_POST['EvilTwin'])){echo date('h:i:s A');EvilTwinLogDate();echo date('h:i:s A');}
elseif(isset($_POST['InsertEvil'])){ echo date('h:i:s A');
InsertEvilTwin($_POST['ZoneInsertEvil'], $_POST['RadioInsertEvil']);echo date('h:i:s A');}
elseif(isset($_POST['CAUMacAddressControl'])){ echo date('h:i:s A');CAUMacAddressControl();echo
date('h:i:s A');}
elseif(isset($_POST['CPEMacAddressControl'])){echo date('h:i:s A');
CPEMacAddressControl($assignedzone); echo date('h:i:s A');}
elseif(isset($_POST['RogueClientMacControl'])){echo date('h:i:s A');
RogueClientMacControl($assignedzone); echo date('h:i:s A'); }
elseif(isset($_POST['RFJamming'])){echo date('h:i:s A');RFJamming();echo date('h:i:s );}
elseif(isset($_POST['CAUSiteSurvey'])){SiteSurveyCAU($assignedzone)}
elseif(isset($_POST['CPESiteSurvey'])){CPESiteSurveys()}
elseif(isset($_POST['LogTheTraffic'])){ echo date('h:i:s A');
LogTraffic($_POST['ZoneLogTraffic'], $_POST['RadioLogTraffic']);echo date('h:i:s A');}
elseif(isset($_POST['MonkeyintheMiddle'])){echo date('h:i:s A');EvilTwin();echo date('h:i:s
A');}else{
print "<BR><BR><form name='GiacForm' method='POST' action='$_SERVER[PHP_SELF]'>";
print "<table align= 'center' border= '1px solid #BDBCBD' border-collapse='collapse'
bgcolor='#0044FF'>";
print "<tr
class='bigsort'><td
colspan='3'><center><font
color='red'
size=6>Wireless
Key fingerprint
= AF19
FA27 2F94 998D
FDB5 DE3D F8B5 06E4
A169 4E46
Rogue Device Detection</font></center></td></tr>";
print "<tr><td><b>CAU Mac Address Control</b></td><td colspan='2'><font
color='blue'><center><input type='submit' name='CAUMacAddressControl' value='CAU Mac
Address Control' size='15'></center></font></td></tr>";
print "<tr><td><b>CPE Mac Address List Control</b></td><td colspan='2'><font
color='blue'><center><input type='submit' name='CPEMacAddressControl' value='CPE Mac
Address List Control' size='15'></center></font></td></tr>";
print "<tr><td><b>Rogue Client Mac Control</b></td><td colspan='2'><font
color='blue'><center><input type='submit' name='RogueClientMacControl' value='Rogue
Client Mac Control' size='15'></center></font></td></tr>";
print "<tr><td><b>Evil Twin</b></td><td colspan='2'><font color='blue'><center><input
type='submit' name='EvilTwin' value='Evil Twin' size='15'></center></font></td></tr>";
print "<tr><td><b>Evil Twin based on Log Date</b></td><td colspan='2'><font
color='blue'><center><input type='submit' name='EvilTwinLogDate' value='EvilTwin Log
Date' size='15'></center></font></td></tr>";
print "<tr><td><b>Monkey in the Middle</b></td><td colspan='2'><font
color='blue'><center><input type='submit' name='MonkeyintheMiddle' value='Monkey in the
Middle' size='15'></center></font></td></tr>";
print "<tr><td><b>RF Jamming</b></td><td colspan='2'><font color='blue'><center><input
type='submit' name='RFJamming' value='RF Jamming' size='15'></center></font></td></tr>";
print "<tr><td><b>Log The Traffic</b></td><td><select name='ZoneLogTraffic'>";
print "<option value='0'>---------------------------";
$sql
= "SELECT * FROM somezones WHERE zone like '84%' OR zone like '32%' OR zone like
'33%' OR zone like '34%' ORDER BY zone";
$r
= mcq($sql,$db);
while($note = mysql_fetch_array($r))
if (strlen($note[zone])>0)
print "<option value='$note[zone]'>$note[zone]";
print "</select>";
Ibrahim Halil Saruhan
© SANS Institute 2007,
52
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
fu
ll r
igh
ts.
print "<select name='RadioLogTraffic'>";
print "<option value=1>1";
print "<option value=2>2";
print "<option value=3>3";
print "</select></td>";
print "<td><font color='blue'><center><input type='submit' name='LogTheTraffic'
value='Log The Traffic' size='15'></center></font></td></tr>";
print "<tr><td><b>CPE Site Survey</b></td><td><select name='Zone'>";
print "<option value='0'>---------------------------";
$sql
= "SELECT * FROM somezones WHERE zone like '84%' OR zone like '32%' OR zone like
'33%' OR zone like '34%' ORDER BY zone";
$r
= mcq($sql,$db);
while($note = mysql_fetch_array($r))
if (strlen($note[zone])>0)
print "<option value='$note[zone]'>$note[zone]";
print "</select></td>";
print "<td><font color='blue'><center><input type='submit' name='CPESiteSurvey'
value='CPE Site Survey' size='15'></center></font></td></tr>";
te
20
Key
07
,A
ut
ho
rr
eta
ins
print "<tr><td><b>CAU Site Survey</b></td><td>";
print "<select name='Zone'>";
print "<option value='0'>---------------------------";
$sql
= "SELECT * FROM somezones WHERE zone like '84%' OR zone like '32%' OR
zone like '33%' OR zone like '34%' ORDER BY zone";
$r
= mcq($sql,$db);
while($note = mysql_fetch_array($r))
if (strlen($note[zone])>0)
print "<option value='$note[zone]'>$note[zone]";
print "</select></td><td><font color='blue'><center><input type='submit'
name='CAUSiteSurvey' value='CAU Site Survey' size='15'></center></font></td></tr>";
print "<tr><td><b>Insert Evil</b></td><td>print "<select name='ZoneInsertEvil'>";
print "<option value='0'>---------------------------";
$sql
= "SELECT * FROM somezones WHERE zone like '84%' OR zone like '32%' OR
zone like '33%' OR zone like '34%' ORDER BY zone";
$r
= mcq($sql,$db);
while($note = mysql_fetch_array($r))
ifFA27
(strlen($note[zone])>0)
fingerprint = AF19
2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
print "<option value='$note[zone]'>$note[zone]";
print "</select>";
In
sti
tu
print "<select name='RadioInsertEvil'>";
print "<option value=1>1";
print "<option value=2>2";
print "<option value=3>3";
print "</select></td>";
print "<td><font color='blue'><center><input type='submit' name='InsertEvil'
value='Insert Evil' size='15'></center></font></td></tr></table></form>";
SA
Log Traffic
NS
}
©
function LogTraffic($zoneid, $radioid){
if ($radioid == 1) $port = 23;
if ($radioid == 2) $port = 24;
if ($radioid == 3) $port = 22;
$result
= mysql_fetch_assoc(mysql_query("SELECT NasIPAddress as CauIpAddress FROM
RadiusNas WHERE (zone = $zoneid)"));
$CauIpAddress = $result[CauIpAddress];
$conn
= ssh2_connect($CauIpAddress,$port);
Ibrahim Halil Saruhan
© SANS Institute 2007,
53
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
if (ssh2_auth_password($conn,'UserName','Password'))
echo "<br><center>Connected to the CAU at $zoneid.</center><br>";
else
die('<br><center>Authentication is Failed...</center><br>');
= '/sbin/iwconfig wlan0 mode monitor;';
date("Y-m-d-H-i-s");
'DUMP' . '_' . $zoneid . '_' . $radioid . '_' . $Now;
'/usr/sbin/tcpdump -i wlan0 -c 3000 -w ' . $dumpcommand . ' ;';
'/bin/sleep 2;';
'/sbin/iwconfig wlan0 mode master;';
'/usr/bin/scp ' . $dumpcommand . ' cpestatus@XX.XX.XXX.XX:.';
fu
ll r
igh
ts.
$iwconfigmonitorcommand
$Now
=
$dumpcommand
=
$tcpdumpcommand
=
$sleepcommand
=
$iwconfigmastercommand =
$scpcommand
=
$allcommands
= $iwconfigmonitorcommand . $tcpdumpcommand . $sleepcommand .
$iwconfigmastercommand . $sleepcommand . $scpcommand ;
$stream
sleep(20);
= ssh2_exec($conn, $allcommands);
ins
}
rr
eta
RF Jamming
function RFJamming(){
ho
= date("Y-m-d H:i:s");
= date("Y-m-d H:i:s",strtotime("$Now -4 hour"));
ut
$Now
$FourHoursAgo
07
,A
$result
= mysql_query("SELECT CPEssid, UserName, CAUNoise FROM logCPE WHERE
LogDate>'$FourHoursAgo' AND NoiseLevel<>0 AND CAUNoise<>0 AND CPEssid<>'' GROUP BY
CPEssid ORDER BY CAUNoise DESC limit 20");
print "<table align= 'center' border= '1px solid #BDBABD' border-collapse='collapse'
AP
List</b></font></td></tr><tr><td><b>Access Point</b></td><td><b>UserName</b></td>
<td><b>CAUNoise</b></td></tr>";
while($note = mysql_fetch_assoc($result))
print "<tr><td><b>$note[CPEssid]</b></td>
<td><b>$note[UserName]</b></td><td><b>$note[CAUNoise]</b></td></tr>";
tu
te
20
bgcolor='#F7F7F7'><tr><td
align='center'
color='red'><b>Noisy
Key fingerprint
= AF19 FA27 2F94
998D FDB5colspan=3><font
DE3D F8B5 06E4
A169 4E46
sti
print "</table><BR><BR>";
NS
In
$result
= mysql_query("SELECT MACAddress, UserName, NoiseLevel FROM logCPE WHERE
LogDate>'$FourHoursAgo' AND NoiseLevel<>0 AND CAUNoise<>0 GROUP BY MACAddress ORDER BY
NoiseLevel DESC limit 20");
SA
print "<table align= 'center' border= '1px solid #BDBABD' border-collapse='collapse'
bgcolor='#F7F7F7'><tr><td align='center' colspan=3><font color='red'><b>Noisy Client
List</b></font></td></tr>";
©
print "<tr><td><b>MAC Address of the Client</b></td><td><b>UserName</b></td>
<td><b>NoiseLevel</b></td></tr>";
while($note = mysql_fetch_assoc($result))
print "<tr><td><b>$note[MACAddress]</b></td><td><b>$note[UserName]</b></td>
<td><b>$note[NoiseLevel]</b></td></tr>";
print "</table>";
}
Client Site Survey
error_reporting(E_ERROR);
Ibrahim Halil Saruhan
© SANS Institute 2007,
54
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
$Connectedarray
= ConnecttoZone($AssignedZoneid, $db);
$conn
= $Connectedarray[0] ;
$Connected
= $Connectedarray[1] ;
if ($Connected == true)
SiteSurveyCPE($conn, $zoneid);
fu
ll r
igh
ts.
function ConnecttoZone($zoneid, $db){
$sql
= "SELECT NasIPAddress FROM RadiusNas WHERE (zone = '$zoneid')";
$result
= mysql_query($sql,$db);
if (!$result) die('Connection Problem:' . mysql_error());
$result
= mysql_fetch_assoc(mysql_query($sql));
$resultIPAddress
= $result[NasIPAddress];
$conn
= ssh2_connect($resultIPAddress,22);
$Connected
= false;
eta
ins
if (ssh2_auth_password($conn,'UserName','Password')){
echo "<br><center>Connected to the CAU.</center><br>";
$Connected = true;
}else{
die('<br><center>Authentication is Failed...</center><br>');
$Connected = false;
}
return array($conn, $Connected);
}
WHERE
ho
rr
function SiteSurveyCPE($conn, $zoneid){
$ListeningIP
= "XX.XX.XXX.XX";
$result
= mysql_query("SELECT id, CPE_IP, zone, CPE_MAC, user FROM Clients
(zone = $zoneid) AND STATUS='enabled' ");
ut
while($note = mysql_fetch_assoc($result)){
07
,A
te
20
Key
$CpeIpAddress = $note[CPE_IP];
$MACAddress
= $note[CPE_MAC];
$CpeZoneId
= $note[zone];
$CustomerId
= $note[id];
WaitOnServer($ListeningIP,
'1',06E4
$CpeIpAddress,
fingerprint$notification
= AF19 FA27=2F94
998D FDB5 DE3D F8B5
A169 4E46$MACAddress,
$conn, '1');
if ($notification == 1)
$filename = LogSiteSurvey($CpeIpAddress, $CpeZoneId, $conn, $CustomerId);
}
tu
}
In
sti
function make_seed() {
list($usec, $sec) = explode(' ', microtime());
return (float) $sec + ((float) $usec * 100000);
}
©
SA
NS
Function WaitOnServer($ListeningIP, $ListeningType, $CpeIpAddress, $MACAddress, $conn,
$Channel){
set_time_limit(40);
$socket = socket_create(AF_INET, SOCK_STREAM, 0) or die("Could not create socket\n");
srand(make_seed());
$ListeningPort = rand();
$ListeningPort = ($ListeningPort % 25)+5501;
while (socket_bind($socket, $ListeningIP, $ListeningPort) != 1) {
$ListeningPort = rand();
$ListeningPort = ($ListeningPort % 25)+5501;
socket_shutdown($socket, 2);
socket_close($socket);
$socket = socket_create(AF_INET, SOCK_STREAM, 0) or die("Could not create
socket\n");
}
$snrcommand
= '/bin/bash -C ' .'\'/mnt/mon/sitesurvey.sh\' \'' .
$CpeIpAddress . '\' \'-p\' \'' . $ListeningPort . '\' \'' . $MACAddress . '\' ' .
' \'&\'';
Ibrahim Halil Saruhan
© SANS Institute 2007,
55
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
fu
ll r
igh
ts.
$stream
= ssh2_exec($conn, $snrcommand);
$result
= socket_listen($socket, 1) or die("Could not set up socket
listener\n");
socket_set_option($socket,SOL_SOCKET, SO_RCVTIMEO, array("sec"=>30,
"usec"=>100));
if ($spawn
= socket_accept($socket)){
$notification = socket_read($spawn,8) or die("Could not read input\n");
if ($notification == 1)
print " <center> CPE Status run successfully!!!</center><BR>";
return($notification);
socket_shutdown($socket, 2);
socket_close($socket);
}else{
socket_shutdown($socket, 2);
socket_close($socket);
print "SOCKET didn't work!!!";
die("Couldn't create socket, error code is: " . socket_last_error() .
",error message is: " . socket_strerror(socket_last_error()));
}
ins
}
ut
ho
rr
eta
function LogSiteSurvey ($CpeIpAddress, $CpeZoneId, $conn, $CustomerId){
$list
= split("\.",$CpeIpAddress);
$XmlCpeIpAddress= $list[2] . $list[3];
$snrstring
= 'scp /var/spool/Client' . $CpeZoneId . '_' . $XmlCpeIpAddress .
'.xml cpestatus@XX.XX.XXX.XX:.';
$stream
= ssh2_exec($conn, $snrstring);
$filename
= '/home/cpestatus/cpe' . $CpeZoneId . '_' . $XmlCpeIpAddress .
'.xml';
07
,A
while (!(file_exists($filename)))
usleep(2000);
XMLParserSiteSurvey($CpeZoneId, $filename, $CustomerId);
return ($filename);
}
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
20
function XMLParserSiteSurvey($CpeZoneId, $filename, $CustomerId){
te
global $AllData, $AddData, $CustId, $CpeZone;
= $CustomerId;
= $CpeZoneId;
sti
tu
$CustId
$CpeZone
NS
In
$parser
= xml_parser_create();
$AllData
= "<table align='center'>";
$AllData
= $AllData . '<tr><td align="center" colspan="8"><b>CPE
INFORMATION</b></td></tr>';
$AddData
= false;
SA
xml_set_element_handler($parser,"startsurvey","stopsurvey");
xml_set_character_data_handler($parser,"proccessingsurvey");
©
sleep(1);
$fp=fopen($filename,"r");
while ($data=fread($fp,4096)){
xml_parse($parser,$data,feof($fp)) or die (sprintf("XML Error: %s at line %d",
xml_error_string(xml_get_error_code($parser)),
xml_get_current_line_number($parser)));
}
$AllData = $AllData . '</tr></table>'; print "$AllData <BR>";
xml_parser_free($parser);
}
Ibrahim Halil Saruhan
© SANS Institute 2007,
56
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
AP Site Survey
SiteSurveyCAU($AssignedZoneId);
function SiteSurveyCAU($zoneid){
= "SELECT NasIPAddress FROM RadiusNas WHERE (zone = $zoneid)";
= mysql_fetch_array(mcq($sql,$db));
$conn
= ssh2_connect($result[NasIPAddress],22);
fu
ll r
igh
ts.
$sql
$result
if (ssh2_auth_password($conn,'UserName','Password'))
echo "<br><center>Connected to the Zone: $zoneid </center><BR>";
else
die('<br><center>Authentication is Failed...</center><br>');
=
=
=
=
=
'/sbin/iwconfig wlan0 mode managed;';
'/bin/sleep 15;';
'/sbin/iwlist wlan0 scan;';
'/sbin/iwconfig wlan0 mode master;';
'cat /proc/net/hostap/wlan0/scan_results;';
$allcommands
$catcommand ;
= $iwconfigmanaged . $sleepcommand . $iwlist . $iwconfigmaster .
$stream
= ssh2_exec($conn, $allcommands);
ut
ho
rr
eta
ins
$iwconfigmanaged
$sleepcommand
$iwlist
$iwconfigmaster
$catcommand
$Dump = '';
07
,A
sleep(20);
$passed = false; $TotalLines = '';
while($line = fgets($stream)){
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
= 'CHID ANL SL BcnInt Capab Rate BSSID ATIM SupRates SSID';
= strpos($line, $findme);
20
$findme
$pos
sti
tu
te
if (!($pos === false))
$passed = true;
if ($passed == true)
ParseLine($line, $zoneid);
In
}
print "</table>";
NS
}
function ParseLine($line, $zoneid){
= ' ';
= strpos($line, $findme);
= substr($line, 0, $pos);
©
SA
$findme
$pos
$channel
if (is_numeric($channel)){
print
"<tr><td>$channel</td>";
$newline
= substr($line, $pos+1);
$pos
$noise
$newline
= strpos($newline, $findme);
= substr($newline, 0, $pos);
= substr($newline, $pos+1);
$pos
$signal
$SNR
= strpos($newline, $findme);
= substr($newline, 0, $pos);
= $signal - $noise;
Ibrahim Halil Saruhan
© SANS Institute 2007,
57
As part of the Information Security Reading Room
Author retains full rights.
"<td>$SNR</td>";
= substr($newline, $pos+1);
$pos
$BcnInt
print
$newline
= strpos($newline, $findme);
= substr($newline, 0, $pos);
"<td>$BcnInt</td>";
= substr($newline, $pos+1);
$pos
$Capab
print
$newline
= strpos($newline, $findme);
= substr($newline, 0, $pos);
"<td>$Capab</td>";
= substr($newline, $pos+1);
$pos
$Rate
print
$newline
= strpos($newline, $findme);
= substr($newline, 0, $pos);
"<td>$Rate</td>";
= substr($newline, $pos+1);
$pos
$mac
print
$newline
= strpos($newline, $findme);
= substr($newline, 0, $pos);
"<td>$mac</td>";
= substr($newline, $pos+1);
$pos
$atim
print
$newline
= strpos($newline, $findme);
= substr($newline, 0, $pos);
"<td>$atim</td>";
= substr($newline, $pos+1);
$pos
$suprates
print
$ssid
= strpos($newline, $findme);
= substr($newline, 0, $pos);
"<td>$suprates</td>";
= substr($newline, $pos+1);
print
"<td>$ssid</td></tr>";
$LogDate
= date("Y-m-d H:i:s");
07
,A
ut
ho
rr
eta
ins
print
$newline
fu
ll r
igh
ts.
Detecting and Preventing Rogue Devices on the Network
Key fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
"<br><table align='center' class='querytable'>";
"<tr>";
"<td>Channel</td>";
"<td>SNR</td>";
"<td>BcnInt</td>";
"<td>Capab</td>";
"<td>Rate</td>";
"<td>Mac</td>";
"<td>ATIM</td>";
"<td>SupRates</td>";
"<td>SSID</td>";
"</tr>";
©
SA
NS
print
print
print
print
print
print
print
print
print
print
print
print
sti
}else{
= mysql_query($sql);
In
$result
tu
te
20
$sql
= "INSERT INTO causitesurveys ";
$sql
= $sql . "(channel, snr, capab, rate, mac, suprates, zoneid, ssid,
LogDate) ";
$sql
= $sql . "VALUES ($channel, $SNR, '$Capab', $Rate, '$mac',
'$suprates', $zoneid, '$ssid', '$LogDate')";
}
}
Ibrahim Halil Saruhan
© SANS Institute 2007,
58
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
Man in the Middle Attack
I decided to detect Evil Twin attack and coded it everywhere like that bu all I
detected was Man-in-the-middle attack. Therefore I changed the subject from Evil
fu
ll r
igh
ts.
Twin to Man-in-the-middle attack.
function EvilTwin(){
$result
= mysql_query("select * from cpesitesurveys where site_ssid like '%BBC%'
OR site_ssid like '%TBZ%' group by mac order by site_ssid");
ins
$previous_ssid = 'none';
$previousmac
= 'none';
rr
eta
print "<table align= 'center' border= '1px solid #BDBABD' border-collapse='collapse'
bgcolor='#F7F7F7'>";
print "<tr><td align='center' colspan=8><b>Different Mac Address Same
SSID</b></td></tr>";
"<tr>";
"<td><b>SSID</b></td>";
"<td><b>$note[site_ssid]</b></td>";
"<td><font color='red'>Mac Address</font></td>";
"<td><b>$note[mac]</b></td>";
"<td><font color='red'>Log Date</font></td>";
"<td><b>$note[LogDate]</b></td>";
"<td><font color='red'>Signal</font></td>";
"<td><b>$note[signal]</b></td>";
"</tr>";
NS
In
sti
tu
print
print
print
print
print
print
print
print
print
print
te
20
Key
07
,A
ut
ho
while($note = mysql_fetch_assoc($result)){
if ($previous_ssid == $note[site_ssid]){
print "<tr>";
print "<td><b>SSID</b></td>";
print "<td><b>$previous_ssid</b></td>";
print "<td><font color='blue'>Mac Address</font></td>";
print "<td><b>$previousmac</b></td>";
print "<td><font color='blue'>Log Date</font></td>";
print "<td><b>$previousLogDate</b></td>";
print
"<td><font
color='blue'>Signal</font></td>";
fingerprint = AF19
FA27
2F94 998D
FDB5 DE3D F8B5 06E4 A169 4E46
print "<td><b>$previoussignal</b></td>";
print "</tr>";
©
SA
}else {
$previous_ssid
$previoussignal
$previousmac
$previousLogDate
=
=
=
=
$note[site_ssid];
$note[signal];
$note[mac];
$note[LogDate];
}
}
print "</table>";
}
function InsertEvilTwin($ZoneId, $Radio){
$Now
$SSID
$tempmac
= date("Y-m-d H:i:s");
= 'BBC_' . $ZoneId . '_' . $Radio;
= 'AA:AA:AA:AA:AA' . ':' . $Radio . $Radio;
Ibrahim Halil Saruhan
© SANS Institute 2007,
59
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
$SQL
= "INSERT INTO cpesitesurveys (zoneid, customerid, site_ssid, mac,
channel, LogDate, signal) VALUES ($ZoneId, 99999, '$SSID','$tempmac', 1, '$Now', -99)";
mysql_query($SQL) OR die(mysql_error());
print"<BR><center><b>Evil Twin Inserted into zone : $ZoneId, radio :
$Radio!</b></center><BR>";
}
fu
ll r
igh
ts.
Evil Twin
This code is similar to the Man-in-the-middle attack but checks log dates and
MAC addresses. This code catches evil twin attack if there is really one. I inserted
ins
data to the DB and saw that it is working, but couldnʼt be able to change the MAC
eta
Address of my USB Wireless Adapterʼs MAC address and couldnʼt be able to test
rr
it on the field.
ho
function EvilTwinLogDate(){
while($note = mysql_fetch_assoc($result)){
$AnHourAgo
= date("Y-m-d H:i:s",strtotime("$Now -1 hour"));
$resultinner
= mysql_query("SELECT * FROM cpesitesurveys WHERE
site_ssid='$note[site_ssid]'
order 4E46
by customerid,
fingerprintLogDate>'$AnHourAgo'
=
AF19
FA27
2F94 AND
998D
FDB5 DE3D F8B5 06E4 A169
LogDate ASC ");
print "<table align= 'center' border= '1px solid #BDBABD' border
collapse='collapse' bgcolor='#F7F7F7'>";
te
20
Key
07
,A
ut
$result
= mysql_query("SELECT site_ssid FROM cpesitesurveys WHERE site_ssid like
'%BBC_8%' OR site_ssid like '%TBZ_8%' group by mac order by site_ssid");
NS
In
sti
tu
print "<tr>";
print "<td><b>Customer Id</b></td>";
print "<td><b>Site SSID</b></td>";
print "<td><b>Channel</b></td>";
print "<td><b>Log Date</b></td>";
print "<td><b>Signal</b></td>";
print "<td><b>Attack Type</b></td>";
print "</tr>";
©
SA
$previouscustomerid
$previousLogDate
$previouschannel
$previousmac
$previoussignal
=
=
=
=
'0';
'0000-00-00 00:00:00';
'0';
'00:00:00:00:00:00';
= '0';
while($noteinner = mysql_fetch_assoc($resultinner)){
if ($previouscustomerid == $noteinner[customerid]){
if ($previousLogDate == $noteinner[LogDate]){
if ($previouschannel == $noteinner[channel])
$AttackType = 'Evil Twin' ;
else
$AttackType = 'Man in The Middle' ;
Ibrahim Halil Saruhan
© SANS Institute 2007,
60
As part of the Information Security Reading Room
Author retains full rights.
Detecting and Preventing Rogue Devices on the Network
if (($previousmac == $noteinner[mac]) && ($previoussignal !=
$noteinner[signal]))
$AttackType = 'Evil Twin' ;
else
$AttackType = 'Man in The Middle' ;
fu
ll r
igh
ts.
print "<tr><td><font color='red'><b>$noteinner[customerid]</b>
</font></td><td><font color='red'><b>$noteinner[site_ssid]
</b></font></td><td><font color='red'><b>$noteinner[channel]</b>
</font></td><td><font color='red'><b>$noteinner[LogDate]</b>
</font></td><td><font color='red'><b>$noteinner[signal]</b>
</font></td><td><font color='red'><b>$AttackType</b>
</font></td></tr>";
}else{
eta
ins
print "<tr><td><font color='green'><b>$noteinner[customerid]</b>
</font></td><td><font color='green'><b>$noteinner[site_ssid]</b>
</font></td><td><font color='green'><b>$noteinner[channel]</b>
</font></td><td><font color='green'><b>$noteinner[LogDate]</b>
</font></td><td><font color='green'><b>$noteinner[signal]</b>
</font></td><td><font color='green'><b>None</b>
</font></td></tr>";
}
}
ho
=
=
=
=
$noteinner[customerid];
$noteinner[LogDate];
$noteinner[channel];
$noteinner[mac];
= $noteinner[signal];
07
,A
}
print "</table>";
ut
$previouscustomerid
$previousLogDate
$previouschannel
$previousmac
$previoussignal
rr
}
©
SA
NS
In
sti
tu
te
20
}
Key
fingerprint = AF19 FA27 2F94 998D FDB5 DE3D F8B5 06E4 A169 4E46
Ibrahim Halil Saruhan
© SANS Institute 2007,
61
As part of the Information Security Reading Room
Author retains full rights.
Last Updated: June 8th, 2017
Upcoming SANS Training
Click Here for a full list of all Upcoming SANS Events by Location
SANS Milan 2017
Milan, IT
Jun 12, 2017 - Jun 17, 2017
Live Event
SANS Charlotte 2017
Charlotte, NCUS
Jun 12, 2017 - Jun 17, 2017
Live Event
SANS Secure Europe 2017
Amsterdam, NL
Jun 12, 2017 - Jun 20, 2017
Live Event
SANS Rocky Mountain 2017
Denver, COUS
Jun 12, 2017 - Jun 17, 2017
Live Event
SANS Minneapolis 2017
Minneapolis, MNUS
Jun 19, 2017 - Jun 24, 2017
Live Event
DFIR Summit & Training 2017
Austin, TXUS
Jun 22, 2017 - Jun 29, 2017
Live Event
SANS Paris 2017
Paris, FR
Jun 26, 2017 - Jul 01, 2017
Live Event
SANS Cyber Defence Canberra 2017
Canberra, AU
Jun 26, 2017 - Jul 08, 2017
Live Event
SANS Columbia, MD 2017
Columbia, MDUS
Jun 26, 2017 - Jul 01, 2017
Live Event
SEC564:Red Team Ops
San Diego, CAUS
Jun 29, 2017 - Jun 30, 2017
Live Event
SANS London July 2017
London, GB
Jul 03, 2017 - Jul 08, 2017
Live Event
Cyber Defence Japan 2017
Tokyo, JP
Jul 05, 2017 - Jul 15, 2017
Live Event
SANS Munich Summer 2017
Munich, DE
Jul 10, 2017 - Jul 15, 2017
Live Event
SANS Los Angeles - Long Beach 2017
Long Beach, CAUS
Jul 10, 2017 - Jul 15, 2017
Live Event
SANS Cyber Defence Singapore 2017
Singapore, SG
Jul 10, 2017 - Jul 15, 2017
Live Event
SANS ICS & Energy-Houston 2017
Houston, TXUS
Jul 10, 2017 - Jul 15, 2017
Live Event
SANSFIRE 2017
Washington, DCUS
Jul 22, 2017 - Jul 29, 2017
Live Event
Security Awareness Summit & Training 2017
Nashville, TNUS
Jul 31, 2017 - Aug 09, 2017
Live Event
SANS San Antonio 2017
San Antonio, TXUS
Aug 06, 2017 - Aug 11, 2017
Live Event
SANS Prague 2017
Prague, CZ
Aug 07, 2017 - Aug 12, 2017
Live Event
SANS Boston 2017
Boston, MAUS
Aug 07, 2017 - Aug 12, 2017
Live Event
SANS Hyderabad 2017
Hyderabad, IN
Aug 07, 2017 - Aug 12, 2017
Live Event
SANS Salt Lake City 2017
Salt Lake City, UTUS
Aug 14, 2017 - Aug 19, 2017
Live Event
SANS New York City 2017
New York City, NYUS
Aug 14, 2017 - Aug 19, 2017
Live Event
SANS Virginia Beach 2017
Virginia Beach, VAUS
Aug 21, 2017 - Sep 01, 2017
Live Event
SANS Chicago 2017
Chicago, ILUS
Aug 21, 2017 - Aug 26, 2017
Live Event
SANS Adelaide 2017
Adelaide, AU
Aug 21, 2017 - Aug 26, 2017
Live Event
SANS Tampa - Clearwater 2017
Clearwater, FLUS
Sep 05, 2017 - Sep 10, 2017
Live Event
SANS San Francisco Fall 2017
San Francisco, CAUS
Sep 05, 2017 - Sep 10, 2017
Live Event
SEC555: SIEM-Tactical Analytics
OnlineCAUS
Jun 12, 2017 - Jun 17, 2017
Live Event
SANS OnDemand
Books & MP3s OnlyUS
Anytime
Self Paced
Open as PDF
Similar pages