Compaq | BL465c - ProLiant - 2 GB RAM | Performance of HP ProLiant BL465c G7 with AMD Opteron 6100

Performance of HP ProLiant BL465c G7
with AMD Opteron 6100 Series
processors in 32- and 64-bit HP SBC
environments
Technical white paper
Table of contents
Executive summary............................................................................................................................... 2
Introduction ......................................................................................................................................... 3
G7 technologies .............................................................................................................................. 3
Test methodology................................................................................................................................. 4
Test tools ......................................................................................................................................... 4
User profile ..................................................................................................................................... 5
Test scenario ................................................................................................................................... 6
Test topology ................................................................................................................................... 7
Configurations ................................................................................................................................. 8
Performance test results....................................................................................................................... 10
Configured with processor Model 6174 ........................................................................................... 11
Configured with processor Model 6164 HE ...................................................................................... 14
Configured with processor Model 6136 ........................................................................................... 17
Configured with processor Model 6128 HE ...................................................................................... 20
Test analysis summary ........................................................................................................................ 22
Comparing scalability in the 32-bit environment ................................................................................ 23
Recommendations .............................................................................................................................. 24
x64 platforms ................................................................................................................................ 24
Avoiding disk I/O bottlenecks ......................................................................................................... 24
Using write cache .......................................................................................................................... 24
Virtualization ................................................................................................................................. 25
Conclusions ................................................................................................................................... 26
Appendix A – Using Microsoft Windows Server 2003 x64 Editions ........................................................ 27
Historical scalability limitations ........................................................................................................ 27
Appendix B – SBC solution sizing ........................................................................................................ 28
Online sizer tool ............................................................................................................................ 28
For more information .......................................................................................................................... 30
Executive summary
Designed for virtualization yet flexible enough for any business application, the HP ProLiant BL465c
G7 server blade combines best-in-class features to help reduce network sprawl and lower
infrastructure costs.
Multi-core AMD Opteron™ 6100 Series processors with up to 16 DIMMs provide the performance
needed for demanding applications and virtualization workloads. The HP ProLiant BL465c G7 server
blade also offers two 10 Gb Ethernet ports with converged network support, two hot plug drive bays
to extend local storage options, and industry-leading management that enables powerful lights-out
administration.
HP tested the optimal scalability of two bare-metal1 HP ProLiant BL465c G7 server blade
configurations, obtaining the following results:
Configuration
AMD Opteron processor
Test
environment
Supported
users
2P/24C2
Model 6174 (2.2 GHz)
64-bit
487
32-bit
141
2P/24C
Model 6164 HE (1.7 GHz)
64-bit
445
2P/16C
Model 6136 (2.4 GHz)
64-bit
411
32-bit
161
64-bit
380
2P/16C
3
Model 6128 HE (2.0 GHz)
Intended audience: This performance characterization is intended primarily for IT professionals
planning HP Server Based Computing (HP SBC) solution deployments. The performance and sizing
information provided herein is designed to help you estimate the number of HP ProLiant BL465c G7
server blades required for a particular environment.
Testing performed in April – May 2010 is described.
1
2
3
2
Non-virtualized
Signifying support for two processors (P) and a total of 12 cores (C)
Where HE denotes high efficiency; this is a lower-powered processor
Introduction
Figure 1 shows the HP ProLiant BL465c G7 server blade.
Figure 1. The HP ProLiant BL465c G7 server blade, illustrating a unique drawer design that allows you to service hard drives
independently
G7 technologies
Seventh-generation AMD Opteron-based HP ProLiant servers such as the BL465c G7 server blade
feature the following key technologies:
 AMD Opteron 6100 Series processors
 DDR3 memory
 Embedded dual-port FlexFabric 10 Gb Ethernet converged network adapter 4
 PCI Express 2.0
 Smart Array controller with 1 GB Flash Backed Write Cache (FBWC)
FBWC
The embedded Smart Array P410i RAID controller5 featured in the HP
ProLiant BL465c G7 server blade provides 1 GB FBWC as standard.
FBWC, which does not require a battery, uses flash memory.
Data retention is indefinite; by comparison, retention with Battery Backed
Write Cache (BBWC) is approximately two days.
4
5
Currently, FlexFabric is only available on HP ProLiant c-Class server blades.
FBWC may be optional on some models. For more information on FBWC, refer to
http://h18004.www1.hp.com/products/quickspecs/13201_div/13201_div.HTML.
3
 HP Integrated Lights-Out 3 (iLO 3)
 Serial Attached SCSI (SAS)
 Thermal logic
The following sections of this paper describe testing performed by HP to characterize the
performance and scalability of an HP ProLiant BL465c G7 server blade in 32- and 64-bit HP SBC
environments.
Test methodology
HP continues to upgrade existing HP ProLiant servers and introduce new servers to meet particular
business needs. To help you select the appropriate server for your particular HP SBC environment, HP
publishes this and other performance characterizations so that you can compare individual server
performance and scalability.
This section describes how HP determined the optimal number of users supported by a 2P HP ProLiant
BL465c G7 server blade featuring a range of AMD Opteron Model 6100 processors – henceforth
referred to as the HP ProLiant BL465c G7 server blade – in 64- and 32-bit test harnesses.
Important:
As with any laboratory testing, the performance metrics quoted in this
paper are idealized. In a production environment, these metrics may be
impacted by a variety of factors.
HP recommends proof-of-concept testing in a non-production environment
using the actual target application as a matter of best practice for all
application deployments. Testing the actual target application in a
test/staging environment identical to, but isolated from, the production
environment is the most effective way to characterize system behavior.
Note:
A 64-bit HP SBC environment eliminates the kernel memory constraints that
can limit server scalability in a 32-bit HP SBC environment. For more
information, refer to Appendix A – Using Microsoft Windows Server 2003
x64 Editions.
This section provides more information on test tools, user profile and test scenarios.
Test tools
To facilitate the placement and management of simulated loads on an HP SBC server, HP used
Terminal Services Scalability Planning Tools (TSScaling), a suite of tools developed by Microsoft® to
help organizations with Microsoft Windows® Server 2003 Terminal Server capacity planning.
Table 1 describes these tools.
4
Table 1. Components of TSScaling
Component
Description
Automation tools
Test tools
Help files
Robosrv.exe
Drives the server-side of the load simulation
Robocli.exe
Helps drive the client-side of the load simulation
Qidle.exe
Determines if any scripts have failed and require
operator intervention
Tbscript.exe
A script interpreter that helps drive the client-side load
simulation
TBScript.doc
Terminal Server bench scripting documentation
TSScalingSetup.doc
A scalability test environment set-up guide
TSScalingTesting.doc
A testing guide
More information
 Roboserver (Robosrv.exe) and Roboclient (Robocli.exe): Terminal Server capacity planning
 TSScaling: Windows Server 2003 Terminal Server Capacity and Scaling
User profile
To simulate a typical workload in the HP SBC test environment, HP used a script based on the
following Heavy User profile:
Heavy Users (also known as Structured Task Workers) tend to open multiple applications
simultaneously and remain active for long periods. Heavy Users often leave applications
open when not in use.
Table 2 outlines the activities performed by Heavy Users, which utilized Microsoft Office 2003
products.
Table 2. Activities incorporated into the test script
6
Activity
Description
Access
Open a database, apply a filter, search through records, add records, and delete
records.
Excel
Open, print and save a large spreadsheet.
InfoPath
Enter data6 into a form; save the form over an existing form.
Outlook
First pass: Email a short message.
Second pass: Email a reply with an attachment.
Outlook_2
Create a long reply.
PowerPoint
Create a new presentation, insert clipart, and apply animation. View the
presentation after each slide is created.
PowerPoint_2
Open and view a large presentation with heavy animation and many colors and
gradients.
Word
Create, save, print, and email a document.
Data entry for Office InfoPath 2003 requires significant processor resources.
5
Test scenario
To characterize its scalability, HP tested the HP ProLiant BL465c G7 server blade in 64- and 32-bit HP
SBC environments when running a workload based on the User profile described above.
Note:
As recommended by HP for most applications, the FBWC was configured
with 100% write for all testing. Use the BIOS to adjust the read/write ratio.
Testing was initiated by running the particular workload with a group of 15 simulated users; start
times were staggered to eliminate authentication overhead. After these sessions finished, HP added
15 more users, then repeated the testing. Further users were added until the optimal number (as
described below) was reached.
Performance and scalability metrics
While the Heavy User workload was running, HP monitored a range of Windows Performance
Monitor (Perfmon) counters in order to characterize the performance and scalability of the tested
server. HP also used canary scripts featuring Office 2003-based activities to establish the number of
users that could be supported before response times became unacceptable.
HP typically uses the Perfmon % Processor Time counter to establish the optimal number of users
supported by an HP SBC server – by definition, the number of users active when processor utilization
reaches 80%. At this time, a limited number of additional users or services can be supported;
however, user response times may become unacceptable.
In a 32-bit HP SBC environment, System Page Table Entries (PTEs) may become exhausted7 before
processor utilization reaches 80%.
To validate metrics obtained from Perfmon, HP uses canary scripts to characterize response times for
a range of discrete activities, such as the time taken to invoke an application or for a modal box to
appear. By monitoring response times – a very practical metric – as more and more users log on, HP
has been able to demonstrate that these times are acceptable when the optimal number of users (as
determined using Perfmon counter values) is active.
Note:
When running canary scripts, HP considers user response times to become
unacceptable when they increase markedly over a baseline measurement.
7
6
Due to inherent limitations in the x86 platform; for more information, refer to Appendix A – Using Microsoft Windows Server 2003 x64
Editions.
Test topology
Figure 2 illustrates the HP SBC test environment.
Figure 2. The tested environment (64-bit implementation shown)
Note:
Test environments such as that shown in Figure 2 are available to customers
at HP Solution Centers to help solve a wide variety of business problems.
7
Configurations
This section outlines the configurations of servers and clients used in the test environment.
Table 3 presents information on system configurations; Table 4 provides a system summary.
Table 3. System configurations
Server
Configuration
HP SBC server
2P HP ProLiant BL465c G7 server blade with:
 AMD Opteron processor Model 6174 (12-core/2.2 GHz), Model 6164 HE (12core/1.7 GHz), Model 6136 (8-core/2.4 GHz), or Model 6128 HE (8-core/2.0
GHz)
– 12 MB shared L3 cache
 64 GB RAM (8 GB for Model 6136, 32-bit testing)
 Smart Array P410i controller with RAID 0
– Two 146 GB 10,000 rpm SAS hard drives
– 96 GB page file on system partition (12 GB for Model 6136, 32-bit testing)
– 1 GB FBWC (100% write)
 HP NC551i Dual Port FlexFabric 10 Gb Ethernet Converged Network Adapter
64-bit: Windows Server 2003 R2 Enterprise x64 Edition with Service Pack 2
32-bit: Windows Server 2003 Enterprise x86 Edition with Service Pack 1
Terminal Services enabled
Office 2003
Exchange Server/
Internet Information Services
2P HP ProLiant DL360 G5 server with:
 Dual-core Intel® Xeon® processor (3.2 GHz)
 2 x 2 MB L2 cache
 2 GB RAM
 Four 72 GB 15,000 rpm SAS hard drives
 Integrated Smart Array P400i controller with RAID 5
 NC373i Multifunction Gigabit Server Adapter
Windows Server 2003 Enterprise Edition
Microsoft Exchange Server 2003
Microsoft Internet Information Services (IIS) 6.0
Continued
8
Table 3. System configurations (continued)
Server
Configuration
Domain controller
2P HP ProLiant DL360 G5 server with:
 Dual-core Intel Xeon processor (3.2 GHz)
 2 x 2 MB L2 cache
 2 GB RAM
 Four 72 GB 15,000 rpm SAS hard drives
 Integrated Smart Array P400i controller with RAID 5
 NC373i Multifunction Gigabit Server Adapter
Windows Server 2003 Enterprise Edition
Client
Variety of Intel Pentium®-based Compaq Evo workstations (600 MHz – 2.533 GHz),
each with:
 At least 256 MB of memory
 1024 x 768/16-bit color depth
 100 Mbps NIC
Windows 2000 Professional or Windows XP
Table 4. System summary for the bare-metal HP ProLiant BL465c G7 server blade (listings shown in bold are unique to the 32bit test environment)
Component
Description
Operating system
Microsoft Windows Server 2003, Enterprise x64 Edition/
Microsoft Windows Server 2003, Enterprise Edition
Version
5.2.3790 Service Pack 2, Build 3790/
5.2.3790 Service Pack 1, Build 3790
Other OS description
R2/Not applicable
System model
ProLiant BL465 G7
System type
x64-based PC/x86-based PC
Processor – each of 24 cores
(Model 6174 or 6164 HE)
or 16 cores (Model 6136 or
6128 HE)
AMD64/x86 Family 16 Model 9 Stepping 1 Authentic AMD
BIOS version/date
HP A19, 1/22/2010
SMBIOS version
2.6
Windows directory
C:\WINDOWS
System directory
C:\WINDOWS\system32
Model
Model
Model
Model
6174: ~2200 MHz
6164 HE: ~1700 MHz
6136: ~2400 MHz
6128 HE: ~2000 MHz
Continued
9
Table 4. System summary for the bare-metal HP ProLiant BL465c G7 server blade (continued)
Component
Description
Boot device
\Device\HarddiskVolume1
Locale
United States
Hardware abstraction layer
5.2.3790.3959 (srv03_sp2_rtm.070216-1710)/
5.2.3790.1830 (srv03_sp1_rtm.050324-1447)
User name
Not available
Total physical memory
Model
Model
Model
Model
6174: 65,531.39 MB/65,531.39 MB
6164 HE: 65,531.39 MB
6136: 65,531.39 MB /8,187.39 MB
6128 HE: 65,531.39 MB
Available physical memory
Model
Model
Model
Model
6174: 61.96 GB/3.25 GB
6164 HE: 62.06 GB
6136: 61.95 GB/3.56 GB
6128 HE: 62.03 GB
Total virtual memory
Model
Model
Model
Model
6174: 158.42 GB/3.25 GB
6164 HE: 158.42 GB
6136: 158.42 GB/3.80 GB
6128 HE:158.42 GB
Available virtual memory
Model
Model
Model
Model
6174: 157.98 GB/3.25 GB
6164 HE: 158.06 GB
6136: 157.97 GB/3.56 GB
6128 HE: 158.04 GB
Page file space
Model
Model
Model
Model
6174: 95.99 GB /95.99 GB
6164 HE: 95.99 GB
6136: 95.99 GB /11.99 GB
6128 HE: 95.99 GB
Page file
C:\pagefile.sys
Performance test results
This section outlines the test results used by HP to characterize the optimal performance and scalability
of the HP ProLiant BL465c G7 server blade. The following metrics were used:
 Perfmon values – Shows select Perfmon counter values for the Heavy User workload
 Canary times – Shows user response times for a sample canary script; used to validate optimal
scalability levels
Note:
As with any laboratory benchmark, the performance metrics quoted in this
performance brief are idealized. In a production environment, these metrics
may be impacted by a variety of factors; for more information, refer to
Appendix B – SBC solution sizing.
HP determined that there were no disk or network bottlenecks in the test
environment.
10
The following scenarios were tested:
 Server blade configured with the AMD Opteron processor Model 6174, 6164 HE, 6136, and
6128 HE
 HP SBC test environments
– 64-bit: 6174, 6164 HE, 6136, 6128 HE
– 32-bit: 6174, 6136
Configured with processor Model 6174
An HP ProLiant BL465c G7 server blade featuring the AMD Opteron processor Model 6174 was
tested in 64- and 32-bit HP SBC test environments.
64-bit test environment
Perfmon values and canary times are presented.
Perfmon values
HP ran a performance test using a workload based on the Heavy User profile. Figure 3 shows select
Perfmon values during this test run; Figure 4 presents normalized results.
Figure 3. % Processor Time values – showing that 487 Heavy Users were supported when processor utilization reached 80%
Figure 3 shows the optimal number of Heavy Users supported by the bare-metal HP ProLiant BL465c
G7 server blade to be 487.
HP noted that, as the test run progressed, processor queue length started to increase steadily and
never emptied. With approximately 500 users logged on (that is, more than the optimal number),
there was a marked increase in queue length.
11
Figure 4 shows normalized Perfmon values.
Figure 4. There were no stopped sessions when 487 Heavy Users were logged on
With 487 users active, there were no stopped sessions.
The processor became saturated with approximately 500 users logged on, which coincided with the
marked increase in processor queue length shown in Figure 3.
The maximum number of users able to log on was 739, though approximately 2% of sessions had
stopped.
12
Canary times
Figure 5 shows sample results for the tested server when running a typical canary script. Individual
user response times are shown in blue, with a yellow line depicting average response times.
HP analyzed this figure to determine when response times began to increase markedly and
consistently over a baseline level, indicating that user response times had become unacceptable.
Figure 5. Canary time values show that user response times started to become unacceptable when 514 Heavy Users were
active
Figure 5 indicates that response times were acceptable when 487 Heavy Users – the optimal number
– were active.
However, more users were able to log on up to a maximum of 739, limited by lack of CPU resources.
32-bit test environment
HP ran a performance test using a workload based on the Heavy User profile. Figure 6 shows select
Perfmon values during this test run.
13
Figure 6. % Processor Time values – showing that 141 Heavy Users were supported when system PTEs became exhausted
The optimal number of Heavy Users supported by the HP ProLiant BL465c G7 server blade was 141.
Configured with processor Model 6164 HE
An HP ProLiant BL465c G7 server blade featuring the AMD Opteron processor Model 6164 HE was
tested in the 64-bit HP SBC test environment.
64-bit test environment
Perfmon values and canary times are presented.
Perfmon values
HP ran a performance test using a workload based on the Heavy User profile. Figure 7 shows select
Perfmon values during this test run; Figure 8 presents normalized results.
14
Figure 7. % Processor Time values – showing that 447 Heavy Users had been able to log on when processor utilization
reached 80%
Thus, as many as 447 Heavy Users were supported when processor utilization reached 80%.
Figure 8 shows normalized Perfmon values.
15
Figure 8. Although 447 Heavy Users were able to log on, two sessions had stopped
With 447 Heavy Users – the number of users logged on when processor utilization reached 80% –
there were two stopped sessions. Thus the optimal number of users is (447 – 2); that is, 445.
16
Canary times
Figure 9 shows sample results for the tested server when running a typical canary script. Individual
user response times are shown in blue, with a yellow line depicting average response times.
HP analyzed this figure to determine when response times began to increase markedly and
consistently over a baseline level, indicating that user response times had become unacceptable.
Figure 9. Canary time values show that user response times started to become unacceptable when 477 Heavy Users were
active
Response times were acceptable when 445 Heavy Users – the optimal number – were active.
Configured with processor Model 6136
An HP ProLiant BL465c G7 server blade featuring the AMD Opteron processor Model 6136 was
tested in 64- and 32-bit HP SBC test environments.
64-bit test environment
Perfmon values and canary times are presented.
Perfmon values
HP ran a performance test using a workload based on the Heavy User profile. Figure 10 shows select
Perfmon values during this test run.
17
Figure 10. % Processor Time values – showing that 411 Heavy Users were supported when processor utilization reached 80%
The optimal number of Heavy Users supported by the bare-metal HP ProLiant BL465c G7 server blade
was 411.
Canary times
Figure 11 shows sample results for the tested server when running a typical canary script. Individual
user response times are shown in blue, with a yellow line depicting average response times.
HP analyzed this figure to determine when response times began to increase markedly and
consistently over a baseline level, indicating that user response times had become unacceptable.
18
Figure 11. Canary time values show that user response times started to become unacceptable when 475 Heavy Users were
active
Response times were acceptable when 411 Heavy Users – the optimal number – were active.
32-bit test environment
HP ran a performance test using a workload based on the Heavy User profile. Figure 12 shows select
Perfmon values during this test run.
19
Figure 12. % Processor Time values – showing that 161 Heavy Users were supported when disk utilization began to increase
exponentially
Disk utilization increased exponentially during this test run due to lack of system PTEs.
The optimal number of users supported by this HP ProLiant BL465c G7 server blade was 161.
Configured with processor Model 6128 HE
An HP ProLiant BL465c G7 server blade featuring the AMD Opteron processor Model 6128 HE was
tested in a 64-bit HP SBC test environment.
64-bit test environment
Perfmon values and canary times are presented.
Perfmon values
HP ran a performance test using a workload based on the Heavy User profile. Figure 13 shows select
Perfmon values during this test run.
20
Figure 13. % Processor Time values – showing that 380 Heavy Users were supported when processor utilization reached 80%
The optimal number of Heavy Users supported by the bare-metal HP ProLiant BL465c G7 server blade
was 380.
Canary times
Figure 14 shows sample results for the tested server when running a typical canary script. Individual
user response times are shown in blue, with a yellow line depicting average response times.
HP analyzed this figure to determine when response times began to increase markedly and
consistently over a baseline level, indicating that user response times had become unacceptable.
21
Figure 14. Canary time values show that user response times started to become unacceptable when 406 Heavy Users were
active
Response times were acceptable when 380 Heavy Users – the optimal number – were active.
Test analysis summary
Figure 15 summarizes the test results.
Figure 15. Optimal numbers of Heavy Users supported by the HP ProLiant BL465c G7 server blade in an HP SBC environment
22
Important:
As with any laboratory benchmark, the performance metrics quoted in this
performance brief are idealized. In a production environment, these metrics
may be impacted by a variety of factors; for more information, refer to
Appendix B – SBC solution sizing.
HP determined that there were no disk or network bottlenecks in the test
environment.
Comparing scalability in the 32-bit environment
Scalability in the 32-bit HP SBC test environment was impacted by the particular processor model
being used, as shown in Figure 16.
Figure 16. More users were supported when the tested server was equipped with 8-core processors
Because the amount of memory consumed by chip-set drivers increases with core density, fewer free
System PTEs were available to the Model 6174 (12-core) processor-powered server at the start of the
test, which translated to support for 38% fewer Heavy Users.
23
Recommendations
This section provides general recommendations for improving the performance of HP SBC servers.
x64 platforms
Since x64 platforms allow you to better utilize memory and multi-core processors, the bottleneck you
are most likely to encounter8 would be associated with the disk subsystem. While a detailed analysis
of disk I/O performance is beyond the scope of this white paper, the following observations are
offered to help you improve disk performance:
 Utilize write cache (see Using write cache)
 Since internal storage is often insufficient to support a large number of users in an HP SBC
environment, consider deploying additional RAID arrays and/or SAN support. Note also that when
a SCSI RAID array is used to host user profiles and page files, the number of spindles deployed has
a significant impact on the response times associated with file access.
When the pressure on the disk I/O subsystem is high, one option for improving disk access times is
to add RAM to lower the pressure on memory.
As the number of sessions increases, disk activity and the pressure on the disk I/O subsystem also
increase. If file I/O activity is high, the probability that requests will find the desired data in
memory decreases, thus negatively affecting file access times.
Avoiding disk I/O bottlenecks
To help you avoid disk I/O bottlenecks, Microsoft recommends using the Windows performance
monitoring tool, Perfmon, to check the following metrics9:
 %Idle time – Idle times for logical and physical drives should average at least 50%
 Average Disk Seconds/Read and Average Disk Seconds/Write – The average time taken to
complete a read or write should average less than 25 milliseconds, with peaks less than 50
milliseconds
If the above conditions specified by Microsoft cannot be met, a disk I/O bottleneck is likely.
Note:
In the event of an I/O bottleneck, you should tune the disk subsystem,
decrease the number of users or applications, or add memory to the server.
Using write cache
HP Smart Array controllers include an allocation of memory that can be utilized to temporarily cache
data being written to or read from disk. Since access to this memory is significantly faster than disk
access, cache can enhance overall server performance, particularly during login operations.
Write cache is of particular interest in an HP SBC environment. After buffering all the data associated
with a particular write command, the Smart Array controller indicates to the HP SBC server that the
data transfer to the disk is complete – even through the data is still being written to disk. This frees up
the server’s processor to perform other tasks and accelerates data throughput.
8
9
24
For further information, refer to the HP white paper, “Scalability and performance of HP ProLiant servers on 64-bit Microsoft
Windows Server 2003 in an HP SBC environment.”
For further information, visit the Microsoft website.
Note:
HP has not yet characterized FBWC performance in the HP SBC
environment. However, testing performed using BBWC demonstrates that,
typically, improvements from write cache are most significant when the HP
SBC server is performing log-intensive operations and/or when significant
page file write operations are necessary, such as during user logins.
Performance gains ranged from 50% to 250%10; actual results would vary
depending on the application(s) involved and your particular HP SBC
environment.
FBWC is enabled by default. Use the BIOS to set the read/write ratio11.
Virtualization
To take best advantage of the benefits delivered by virtualization, you need to understand your HP
SBC environment. In addition, be aware that virtual machine (VM) performance may vary depending
on the application, the guest operating system, and other factors; you should test the VM prior to
implementation in a production environment. Also be aware that some applications are good
candidates for virtualization, others less so: for example, underutilized HP SBC servers, servers
running infrastructure services, and Citrix XenApp data store servers may be good candidates for
virtualization; HP SBC servers running resource-intensive applications and highly-utilized infrastructure
servers may not be such good candidates.
When you are planning a virtualized implementation, ensure you have selected a server with the CPU
capacity and number of cores you need. Make sure you meet the resource requirements for
virtualization overhead, guest operating systems, and applications. Would a storage array network
(SAN) be a better choice than internal storage? Are there enough network interface cards (NICs)?
To correctly size the HP SBC servers you intend to virtualize, you must understand the associated
applications and the numbers of users and user profiles to be supported. You should balance the
number of VMs deployed on a particular server with the number of vCPUs allocated to each VM; you
should also allocate enough RAM to eliminate memory and I/O bottlenecks. Prior to deployment, HP
recommends testing your VMs in a production test environment with live users.
In addition to suitable sizing, optimal VM performance requires Citrix XenServer and guest operating
systems to be appropriately configured. For example, consider disabling the screen saver associated
with the XenServer VM controller window.
To avoid spikes in processor utilization, ensure your VMs are online before applying the workload.
Do not simultaneously add large numbers of users; if possible, balance the workload across your
VMs.
A broad range of tools is available to help you manage a virtualized HP SBC environment, including
Citrix Essentials for XenServer Solution - HP Edition and Citrix XenCenter.
VMs are flexible, allowing you to readily implement the level of availability you need. Moreover, you
can enhance availability by utilizing a storage array network (SAN) created from HP StorageWorks
product offerings. Capabilities may include:
 Multiple paths for redundancy
 Automatic path failover
 High-availability cluster support
10
11
For further information, visit the HP website.
For most applications, HP recommends 100% write.
25
Note:
For more information on best practices associated with virtualization, refer
to the HP white paper, “Best practices for deploying Citrix XenApp on
XenServer.”
Conclusions
When planning an x64 HP SBC environment, you should select servers equipped with multi-core
processors to help maximize scalability. If your budget allows, consider the fastest processors, the
most cores, and largest cache; if your objective is to reduce overall power consumption and you are
prepared to accept a small performance penalty, use low-power processors.
Memory is an important factor in the x64 environment: while an x64 platform can utilize more RAM,
it also has a higher minimum RAM requirement than an x86 platform. Since a system that is not
memory-starved is less likely to experience disk I/O bottlenecks, HP recommends adding as much
RAM as your budget permits.
For optimal performance, configure FBWC for 100% write. Consider using 15,000 rpm SAS drives.
Virtualized server
Do not oversubscribe vCPUs – after a certain point, additional vCPUs can degrade performance.
Determine the optimal server configuration for your particular workload: How many VMs? How many
vCPUs? How much memory? Significant testing may be required.
Consider virtualizing later x86 platforms to take advantage of improvements made to processors,
memory, and I/O. You may experience dramatic scalability enhancements.
26
Appendix A – Using Microsoft Windows Server 2003 x64
Editions
Microsoft offers operating systems that support both 64- and 32-bit applications, as well as existing
32-bit deployment and management tools – all on the same platform. These operating systems
provide an evolutionary path to 64-bit technology, allowing 64- and 32-bit applications to run sideby-side during the gradual migration to 64-bit computing.
64-bit editions of Windows Server 2003 running on today’s multi-core AMD Opteron or Intel Xeon
processors can improve the performance of HP SBC servers by processing more data per clock cycle,
addressing more memory, and running some numerical calculations faster. Large data sets can be
loaded entirely into memory, reducing the need for slower disk access; complex calculations that take
hours to complete on a 32-bit system can be performed in minutes; and workloads that once required
a large server farm can be performed by a single server.
In addition, 64-bit platforms also remove many of the limitations that have previously inhibited
scalability in an HP SBC environment.
Historical scalability limitations
32-bit Windows operating systems can directly address 4 GB of memory, 2 GB of which is reserved
for the operating system kernel and 2 GB for applications. Since kernel memory is shared by all
applications, the relatively small size of this space can be particularly problematic in an HP SBC
environment where a server may be responsible for hundreds of users and thousands of processes. In
this scenario, kernel memory can become constrained, making user response times unacceptably long
and effectively limiting the ability of the server to scale up.
Historically, HP SBC environments have been implemented using 1P or 2P servers. Larger, more
powerful servers have typically not been deployed for two main reasons:
 Kernel memory issues have limited the performance of more powerful servers; either a disk I/O
bottleneck occurs or kernel memory is consumed before processor resources can be fully utilized
 Scalability in a 32-bit symmetric multi-processing (SMP) system is inherently non-linear above 2P
With these 1P and 2P server farms, opportunities to scale up are limited. As a result, customers are
forced to scale out, which can create new problems such as deployment and management
complexity, high power and cooling requirements, under-utilized resources, and minimal opportunities
for server consolidation.
The 64-bit platform shatters the earlier 4 GB limitation – for example, Windows Server 2003 R2
Datacenter x64 Edition with Service Pack (SP) 1 supports up to 2 TB of RAM – effectively removing
kernel memory limitations and eliminating disk I/O bottlenecks. By deploying a Windows Server
x64 Edition operating system, customers can fully utilize the resources of their existing HP SBC
servers and take full advantage of new, more powerful systems – whether they are running 32- or
64-bit applications.
More information
For more information on the impact of 64-bit Windows Server 2003 x64 Editions in an HP SBC
environment, refer to the HP white paper, “Scalability and performance of HP ProLiant servers on 64bit Microsoft Windows Server 2003 in an HP SBC environment.”
To learn about 64-bit computing in an HP SBC environment, refer to the HP white paper,
“Fundamentals of 64-bit computing in an HP SBC environment.”
27
Appendix B – SBC solution sizing
As with any laboratory benchmark, the performance metrics quoted in this performance brief are
idealized. In a production environment, these metrics may be impacted by a variety of factors,
including the following:
 Overhead
Agents and services (virus scanning, backup and restore, provisioning, security, management and
more) automatically consume overhead. Rogue applications can consume additional overhead.
The system architect may wish to provide a 25% – 30% buffer to accommodate this overhead.
 Future growth
To accommodate future growth, the system architect may wish to provide an additional buffer.
Alternatively, servers can be added as needed, taking advantage of the server farm’s inherent
ability to scale out.
 User profiles
The particular application in use directly impacts the number of users supported by a particular
server. Further, user behavior can also impact scalability:
– Increased typing rates correspond to fewer users.
– Opening and closing applications (rather than switching between them) or moving quickly
between tasks can place a heavier load on the server.
– Sizing for this performance test was based on the Heavy User profile described in User profile. If
this profile does not match your needs, more profiles are available using the online sizer tool
(described below); alternatively, the system architect can consult HP Services for more
information.
 Background grammar checking
Background grammar checking can significantly impact scalability, reducing the number of users
supported by as much as 50%. HP disabled background grammar checking for the testing
described in this performance brief.
Online sizer tool
To minimize risk, HP offers automated, downloadable tools that can help you size an HP SBC
solution. A sizer (HP ProLiant Sizer for Citrix XenApp and Microsoft Windows Server 2003 Terminal
Services) is available for enterprise and small and medium business (SMB) environments.
The algorithms and methodology used by this sizer are based on the results of customer surveys and
thorough testing.
Figure B-1 shows the home page for the sizer.
28
Figure B-1. Home page for the HP Citrix XenApp and Microsoft Terminal Services Sizing Tool
Based on information provided by the customer, the sizer can provide a quick, consistent mechanism
for identifying the “best-fit” server for a particular HP SBC environment and generate a Bill of
Materials (BOM) for that server.
29
For more information
HP ProLiant BL465c G7 server blade
http://www.hp.com/servers/bl465c
HP ActiveAnswers solution area for HP SBC,
including Citrix XenApp and Microsoft
Terminal Services
http://www.hp.com/solutions/activeanswers/hpsbc
HP ProLiant Sizer for Citrix XenApp and
Microsoft Windows Server 2003 Terminal
Services
http://h71019.www7.hp.com/ActiveAnswers/cache
/70245-0-0-0-121.html
HP Services
http://www.hp.com/hps/
HP Solution Centers
http://www.hp.com/go/solutioncenters
Citrix XenApp
http://www.citrix.com/site/PS/products/feature.asp?f
amilyID=19&productID=186&featureID=4110
Citrix XenServer
http://h71019.www7.hp.com/ActiveAnswers/cache
/457122-0-0-225-121.html
http://www.citrix.com/English/ps2/products/product
.asp?contentID=683148
Citrix Essentials for XenServer Solution - HP
Edition
http://h18004.www1.hp.com/products/servers/soft
ware/citrix/
To help us improve our documents, please provide feedback at
http://h20219.www2.hp.com/ActiveAnswers/us/en/solutions/technical_tools_feedback.html.
© Copyright 2010 Hewlett-Packard Development Company, L.P. The information contained herein is subject to
change without notice. The only warranties for HP products and services are set forth in the express warranty
statements accompanying such products and services. Nothing herein should be construed as constituting an
additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein.
Microsoft and Windows are U.S. registered trademarks of Microsoft Corporation. Intel, Xeon, and Pentium are
trademarks of Intel Corporation in the U.S. and other countries. AMD Opteron is a trademark of Advanced
Micro Devices, Inc
4AA1-7363ENW, Created June 2010
Download PDF