HP BL35p - ProLiant - 2 GB RAM User manual

Scalability and performance of HP ProLiant BL35p server
blades with AMD Opteron Model 250 (2.4 GHz/1 MB)
processors in an HP SBC environment
Executive summary............................................................................................................................... 3
Introduction......................................................................................................................................... 3
This performance brief...................................................................................................................... 3
Overview............................................................................................................................................ 4
Benefits........................................................................................................................................... 4
Ideal environment ............................................................................................................................ 5
Features.......................................................................................................................................... 5
Performance testing.............................................................................................................................. 6
Test tools and scripts......................................................................................................................... 6
Tools........................................................................................................................................... 6
More information on the test tools................................................................................................... 7
Test scripts................................................................................................................................... 7
Test scenarios .................................................................................................................................. 8
Monitoring processor utilization ..................................................................................................... 8
Validation using a canary script..................................................................................................... 8
Tested environment .............................................................................................................................. 9
Guidelines..................................................................................................................................... 10
Configurations ............................................................................................................................... 11
System summary ............................................................................................................................ 12
Performance test results....................................................................................................................... 13
% Processor Time ........................................................................................................................... 14
Canary time .................................................................................................................................. 15
Results summary............................................................................................................................. 16
Summary .......................................................................................................................................... 16

Appendix A – AMD Opteron processors .............................................................................................. 17
Features........................................................................................................................................ 17
Benefits......................................................................................................................................... 18
Breaking the 4 GB address limit....................................................................................................... 18
Appendix B – SBC solution sizing........................................................................................................ 20
Online sizer tool ............................................................................................................................ 20
For more information.......................................................................................................................... 22

Executive summary
The new HP ProLiant BL35p two-way server blade delivers uncompromising manageability, maximum
computing density, and breakthrough power efficiencies to the high-performance data center. Offering
customers a modular, space-saving design, the HP ProLiant BL35p server blade consumes less power,
enabling denser rack architectures.
The HP ProLiant BL35p shares the same infrastructure components as all BL p-Class server blades, allowing
customers to enjoy additional benefits from the HP BladeSystem p-Class rack-centralized power sub-system
and network interconnect options.
A two-way HP ProLiant BL35p server blade with 2.4 GHz/1 MB processors can provide optimal
support for up to 123 users1in an HP SBC environment.
Introduction
HP offers a broad range of HP ProLiant servers that are ideal for particular HP SBC environments. To
help customers select the appropriate server for their environment, HP publishes a series of
performance briefs that characterize individual server performance and scalability. Each performance
brief includes the following information:
•An outline of the benefits and features introduced with the new server
•A description of the test environment and test scenarios used to determine the optimal number of
users supported by the particular server in an HP SBC environment
HP also publishes a companion white paper that serves as a high-level summary of overall HP
ProLiant server performance and scalability in an HP SBC environment.
This performance brief
This performance brief focuses on the two-way HP ProLiant BL35p server blade with 2.4 GHz/1 MB
processors, providing information on the following topics:
•Overview – Summarizes the benefits and key features of the HP ProLiant BL35p server blade
•Performance testing – Outlines the test tools and scripts implemented by HP to simulate an HP SBC
environment; describes the Heavy, Medium, and Light User profiles used to create the appropriate
workloads
•Tested environment – Provides more information on the HP SBC environment used to test this server
•Performance test results – Provides sample test results; summarizes the optimal numbers of users
supported
•Summary
Note:
HP refers to an application server in an HP SBC environment as an
HP SBC server. Microsoft refers to this server as a Terminal Server;
Citrix as a MetaFrame Presentation Server.
1Heavy Users, as defined later in this Performance Brief
3

Overview
This section outlines the benefits and key features of the HP ProLiant BL35p server blade, shown in
Figure 1.
Figure 1: HP ProLiant BL35p server blade, showing a single blade and two blades in an HP BladeSystem p-Class
Blade Sleeve
Benefits
HP introduces the new HP ProLiant BL35p server blade, which delivers maximum computing density
and breakthrough power efficiencies to the enterprise data center or high-performance technical
computing environments. Benefits include:
Price/performance
•Up to two AMD Opteron™ 200 Series processors running at 2.4GHz, each with 1 MB L2 cache
•Up to 8 GB of PC3200 DDR (400 MHz) memory
Density and power efficiency
•Ultra-dense design supports up to 96 servers in an industry-standard rack
•Lower power consumption per server (68 W per processor)
Management
•HP Integrated Lights-Out (iLO) Advanced standard
•Support for HP ProLiant Essentials Rapid Deployment Pack (RDP)
Options
•Fibre Channel support for SAN implementations and clustering
4

Ideal environment
Medium to large enterprise data center customers requiring dense two-way server performance
•Maximum 32-bit or 64-bit two-way performance for load-balanced or front-end applications
•Ultra-dense design supports up to 96 servers (192 processors) in a standard rack
•Lower power consumption per server (68 W per processor) enables denser rack architectures
High Performance Technical Computing (HPC) environments
•Support for up to two Opteron 200 Series processors with on-board memory controllers running at
processor speed
•Optimal solution for applications that require high bandwidth and low-latency access to system
memory
•Typical applications: financial, life and materials sciences, mechanical engineering/virtual
prototyping, electronic design automation, and high-end film and video
Enterprise data center or HPC environments that optimize external storage architectures
•Dual-port Fibre Channel connectivity option with no trade-offs on other features
•Support for HP and third-party SANs
•Support for boot from SAN
Features
Table 1 summarizes key features of the HP ProLiant BL35p server blade.
Table 1: Key features
Component Description
Processor Opteron Model 250 (2.4 GHz)
Support for one-way and two-way server configurations
1 MB L2 cache per processor
AMD HyperTransport™ (HT) speed: 800 MHz
Up to 192 processors in a 42U rack
Chipset Opteron 8000 Series
Memory ECC PC3200 (400 MHz) DDR
Expandable to 8 GB
Network Two NC7781 PCI-X Gigabit NICs (deployed on a Mezzanine card)
One 10/100 NIC dedicated to remote management
Internal storage Integrated storage controller
Maximum 120 GB (2 x 60 GB ATA 5,400 rpm drive)
SAN connectivity Optional Dual Port Fibre Channel Adapter (2 Gb)
Remote
management Integrated Lights-Out (iLO) Advanced
5

Performance testing
HP continues to upgrade existing HP ProLiant servers and introduce new servers to meet particular
business needs. To help customers select the appropriate server for their particular HP SBC
environment, HP publishes this and other performance briefs to characterize individual server
performance and scalability.
This section describes the test environment and test scenarios used to determine the optimal number of
users supported by an HP ProLiant BL35p server blade in an HP SBC environment.
Note:
HP makes a distinction between the optimal number of users that
can be supported on a particular server configuration and the
maximum number. The optimal number is a more practical metric
that reflects the number of users that can be supported without
impacting user response times.
Historically, HP has defined the optimal value as the number of
users that are active when processor utilization reaches 80%. After
processor utilization passes 80%, additional users are supported;
however, response times may become unacceptable.
Scalability metrics specified in this Performance Brief reflect the
optimal number of users.
Test tools and scripts
HP used a range of test tools and scripts to simulate typical HP SBC user workloads.
Tools
Terminal Services Scalability Planning Tools (TSScaling), a suite of tools developed by Microsoft®,
helps organizations with Microsoft Windows® Server 2003 Terminal Server capacity planning. These
tools facilitate the placement and management of simulated loads on an HP SBC server.
Table 2 outlines the components of the TSScaling suite.
Table 2: Components of TSScaling
TSScaling suite component Description
Robosrv.exe Drives the server-side of the load simulationAutomation tools
Robocli.exe Helps drive the client-side of the load simulation
Qidle.exe Determines if any scripts have failed and require operator
intervention
Test tools
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
Help files
TSScalingTesting.doc A testing guide
6

More information on the test tools
•Roboserver (Robosrv.exe) and Roboclient (Robocli.exe): Terminal Server capacity planning
•TSScaling: Windows Server 2003 Terminal Server Capacity and Scaling
Test scripts
HP used test scripts that simulated the activities of three types of users (Heavy, Medium, and Light).
Table 3 describes these users and their simulated activities.
Table 3: Simulated user activities
User Type Activities
Heavy User 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.
Heavy users have an average typing speed of 40 – 50 words per minute.
The Heavy User script included the following activities:
•Creating, saving and printing documents using Microsoft Word
•Browsing web pages, and sending and receiving e-mail using Microsoft Outlook
•Performing extensive Microsoft Excel activity (using PivotTable
dynamic views, graphing
and charting data, and so on)
Medium User Medium Users (also known as Knowledge Workers) are defined as users who gather,
add value to, and communicate information in a decision-support process. Cost of
downtime is variable but highly visible. These resources are driven by projects and ad-
hoc needs towards flexible tasks. These workers make their own decisions on what to
work on and how to accomplish the task.
Sample tasks include: marketing, project management, sales, desktop publishing,
decision support, data mining, financial analysis, executive and supervisory
management, design, and authoring.
Medium Users tend to open and close applications more quickly than Heavy Users. These
users typically do not leave applications open when not in use; they have an average
typing speed of 35 – 40 words per minute.
The Medium User script included the following activities:
•Creating small documents with Microsoft Word
•Performing simple data entry with Microsoft Excel
•Sending and receiving e-mail with Microsoft Outlook
•Engaging in minimal browsing of an intranet site using Microsoft Internet Explorer 6.0
Light User Light Users (also known as Data Entry Workers) input data into computer systems.
Activities include transcription, typing, order entry, clerical work and manufacturing.
The Light User script simulated simple Microsoft Excel data entry, entering 10 rows of
data.
7

Test scenarios
HP deployed the configuration shown in Figure 2 to simulate an HP SBC environment. To simulate
typical HP SBC workloads, HP ran a series of performance tests based on the Heavy, Medium, and
Light User scripts described in Table 3.
For each test scenario, HP began by running the appropriate script with a group of ten simulated
users. Start times were staggered to eliminate authentication overhead. After the sessions finished, HP
added ten more users, then repeated the testing.
Monitoring processor utilization
Primarily, HP monitored processor utilization to establish the optimal number of users supported by the
HP SBC server. By definition, the optimal number of users is active when processor utilization reaches
80%2.
To obtain this key performance metric, HP used the Windows Performance Monitor (Perfmon) analysis
tool to monitor % Processor Time values.
Validation using a canary script
To validate the scalability metrics obtained using % Processor Time, HP also ran canary scripts to
characterize Heavy User response times for discrete activities (such as the time taken for an
application to be invoked or for a modal box to appear).
By monitoring these response times as more and more users logged on, HP was able to obtain further
scalability metrics.
Note:
When using canary scripts, HP considers optimal user scalability to
be reached when response times increase markedly over a
baseline measurement.
2Historically, HP has defined the optimal number of users as the number of users that are active when processor utilization (%
Processor Time) reaches 80%. Additional users are supported but response times may become unacceptable.
8

Tested environment
Figure 2 illustrates the HP SBC test environment.
Figure 2: The tested environment
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.
9

Guidelines
HP followed these guidelines when configuring servers for the test environment:
•Memory
Since all applications reside and are executed on the HP SBC server, memory resources on this
server must be sufficient in order to optimize system performance. HP offers the following
recommendations based on user type:
–5 – 6 MB for each active Light User
–8 – 10 MB for each active Medium User
–12 – 15 MB for each active Heavy User
For more information on these user types, refer to the section on Test Scripts.
•NIC speed
Since network traffic is light (only keystrokes, mouse clicks and display updates), HP set all server
NICs to 100 Mbps.
10

Configurations
Table 4 summarizes the configurations of systems used in the test environment.
Table 4: System configurations
Server Configuration
HP SBC server Two-way HP ProLiant BL35p server blade with:
•2.4 GHz/1 MB AMD Opteron processors
•1 MB L2 cache
•4096 MB RAM
•One 60.0 GB 5,400 rpm ATA hard drive
Microsoft Windows Server 2003
Enterprise Edition, Terminal Services enabled
Microsoft Office XP
Exchange Server/
Internet Information
Services
HP ProLiant DL380 G2 server with:
•Two-way 1.4 GHz Intel® Pentium® III processor
•512 KB L2 cache
•2304 MB RAM
•Two hot plug 36.4 GB 15,000 rpm Ultra3 SCSI hard drives
•Integrated Smart Array 5i controller with RAID 1
•NC3163 10/100 Fast Ethernet NIC
Microsoft Windows 2000 Advanced Server
Microsoft Exchange Server 2000
Microsoft Internet Information Services 5.0
HP ProLiant DL380 G2 server with:
•Two-way 1.4 GHz Pentium III processor
•512 KB L2 cache
•2304 MB RAM
•Two hot plug 36.4 GB 15,000 rpm Ultra3 SCSI hard drives
•Integrated Smart Array 5i controller with RAID 0
•NC3163 10/100 Fast Ethernet NIC
Domain Controller
Microsoft Windows 2000 Advanced Server
Variety of Pentium-based 600 MHz – 2.533 GHz
Compaq Evo workstations, each with:
•At least 128 MB of memory
•640×480/256-color resolution
•100 Mbps NIC
Client
Microsoft Windows 2000 Professional
11

System summary
Table 5 summarizes the configuration of the HP SBC server.
Table 5: HP SBC server configuration
Component Description
Operating system Microsoft Windows Server 2003, Enterprise Edition
Version 5.2.3790 Build 3790
System name BL35p
System manufacturer HP
System model ProLiant BL35p G1
System type x86-based PC
Processor (each) x86 Family 15 Model 5 Stepping 10 AuthenticAMD ~2405 MHz
BIOS version/date HP A03, 2/15/2005
SMBIOS version 2.3
Windows directory C:\WINDOWS
System directory C:\WINDOWS\system32
Boot device \Device\HarddiskVolume1
Locale United States
Hardware abstraction
layer version 5.2.3790.0 (srv03_rtm.030324-2048)
User name Not available
Time zone Central Standard Time
Total physical memory 4,096.00 MB
Available physical
memory 3.54 GB
Total virtual memory 16.50 GB
Available virtual
memory 16.08 GB
Page file space 12.66 GB
Page file C:\pagefile.sys
12

Performance test results
HP tested a two-way HP ProLiant BL35p server blade with 2.4 GHz/1 MB Opteron processors to
determine the optimal number of users supported in an HP SBC environment.
This section provides information on the following topics:
•% Processor Time – Shows % Processor Time values for a typical test scenario – the Heavy User
scenario in this example
•Canary script response times – Shows response times for a sample canary script that used Excel
charting functions
•Summary – Summarizes test results for Heavy, Medium, and Light Users
Note:
HP determined that there were no disk, memory, or network
bottlenecks in the test environment.
13

% Processor Time
Figure 3 shows % Processor Time values for Heavy Users.
Figure 3: % Processor Time values for Heavy Users – showing an optimal level of 123 users
Figure 3 indicates that an HP ProLiant BL35p server blade can support 123 Heavy Users before the
% Processor Time value reaches 80%.
14

Canary time
Figures 4 shows sample results for an HP ProLiant BL35p server blade running a typical canary script.
Individual user response times are shown in blue, with a yellow line depicting average response
times.
HP analyzed Figure 4 to determine when response times began to increase markedly and consistently
over a baseline level, providing an indication of optimal scalability.
Figure 4: Canary time values – showing an optimal level of 141 Heavy Users
Figure 4 indicates that an HP ProLiant BL35p server blade can support 141 Heavy Users before
response times start to increase significantly, validating the value of 123 Heavy Users derived
using % Processor Time.
15

Results summary
Figure 5 summarizes the optimal numbers of users supported by an HP ProLiant BL35p server blade
(based on % Processor Time values).
Figure 5: Optimal numbers of users supported
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.
Summary
To characterize the performance of the HP ProLiant BL35p server blade, HP selected a metric that
relates directly to the user experience – the number of Heavy Users that can be supported before
response times start to become unacceptable. Test results show that a two-way HP ProLiant BL35p
server blade with Opteron 2.4 GHz/1 MB processors can support 123 Heavy Users, making this
system ideal for an enterprise HP SBC environment.
Customers can take advantage of test environments like the one described in this Performance Brief –
and solve many other business problems – at over 80 strategically-located HP Solution Centers.
16

Appendix A – AMD Opteron processors
The AMD Opteron processors deployed in the latest HP ProLiant servers help advance the future of
industry-standard servers. Key benefits include:
•Improving the performance of existing 32-bit applications
•Supporting 64-bit addressing – breaking the 4 GB limit inherent in earlier x86 processor
implementations
By supporting simultaneous 32- and 64-bit processing, the Opteron processor provides a bridge to
full 64-bit computing.
Features
Figure A-1 illustrates key features of the Opteron architecture
Figure A-1: Key features of the Opteron architecture
17

Benefits
Table A-1 outlines key benefits of the Opteron processor.
Table A-1: Key benefits of the Opteron processor
Feature Benefit
Simultaneous 32- and 64-bit computing
capabilities
Users can run 32-bit and/or 64-bit applications and
operating systems – without sacrificing performance.
Direct Connect Architecture helps reduce
performance bottlenecks
Direct Connect Architecture can reduce memory latency,
provide better balance for throughput, and support more
linear symmetrical multiprocessing.
Support for up to three coherent
HyperTransport links, providing up to 19.2
GB/s peak bandwidth per processor
HyperTransport links provide substantial I/O bandwidth for
current and future application needs.
256 Terabytes (TB) of memory address space More memory space creates a significant performance benefit
for applications where large datasets – or a large number of
datasets – are stored in memory.
Scales from one-way to eight-way The scalability of the processor helps maximize flexibility in
the IT infrastructure.
Integrated memory controller reduces
latencies during memory access
Reduced memory latency translates to increased performance
and productivity.
Low-power processor models are available
•HE – 55 W
•EE – 30 W
Low-power processor models can increase server density and
reduce the Total Cost of Ownership (TCO) for data centers
with limited power budgets. For example, more servers can
be supported at a data center with utility power constraints.
Breaking the 4 GB address limit
Opteron processors take advantage of the 64-bit addressing capabilities of Windows Server
2003 to break the 4 GB direct-addressing limit inherent in earlier x86 implementations.
Note:
Windows Server 2003 for 64-Bit Extended Systems is required to
extend memory space for x86 processors featuring 64-bit
extension technologies.
Information on the release of this operating system is available on
the Microsoft website.
18

Figure A-2 shows the approximate amount of directly-addressable memory available with various
Windows operating systems.
Figure A-2: Memory limits with Windows operating systems
While 32-bit Windows operating systems can address a maximum of 4 GB directly, memory space
can be extended with certain operating systems through the use of Address Windowing Extensions
(AWE).
Windows Server 2003 introduced 64-bit addressing, shattering the earlier 4 GB direct-addressing
limitation. Windows Server 2003 Datacenter Edition can address 512 GB directly – and up to
256 TB indirectly.
19

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.
–For accurate sizing, system architects must closely match their user profiles with the Heavy,
Medium, and Light User profiles used by HP during performance testing. If the profiles do not
match, more 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, online tools that can help the customer size an HP SBC
solution. The algorithms and methodology used by the sizer are based on the results of customer
surveys and thorough testing.
Sizers are available for the following HP SBC environments:
•Enterprise
•Small and Medium Business (SMB)
Figure B-1 shows a typical HP SBC sizer screen.
20