Specifications

ManualsBrandsDell ManualsServer2950

TEST REPORT

FEBRUARY 2006

Virtualized Exchange workload performance comparison of end-to-end

solutions with iSCSI storage connections: Dell PowerEdge 2950 with Dell

EqualLogic PS5000 vs. HP ProLiant DL385 with HP StorageWorks EVA 4400

TEST REPORT

OCTOBER 2008

KEY FINDINGS

z The Dell PowerEdge 2950 server and Dell

EqualLogic PS5000 storage array delivered

21.4 percent more virtualized Exchange

workload performance than the HP ProLiant

DL385 server and HP StorageWorks EVA

4400 storage array (see Figure 1).

z The Dell PowerEdge 2950 server and Dell

EqualLogic PS5000 storage array delivered

18.2 percent more virtualized Exchange

workload performance/watt than the HP

ProLiant DL385 server and HP

StorageWorks EVA 4400 storage array (see

Figure 2).

Executive summary

Dell Inc. (Dell) commissioned Principled Technologies

(PT) to compare the virtualized Exchange 2007 workload

performance of two server-and-storage solutions with

iSCSI connections:

• Dell PowerEdge 2950 server and Dell EqualLogic

PS5000 storage array

• HP ProLiant DL385 server and HP StorageWorks

EVA 4400 storage array

Both servers contained two processors and 32 GB of

RAM, and both connected to their storage with iSCSI.

Each server ran multiple Exchange 2007 workloads, one

per virtual machine (VM). Each VM’s workload used the

Microsoft Exchange Load Generator. By combining

multiple virtual machines with a disk-intensive workload,

the test placed a heavy load on both the server and the

storage. Thus, it measured the performance of the overall solution--server and storage.

Figure 1 compares the peak number of virtual machines running a Microsoft Exchange 2007 workload that each

solution ran with acceptable performance (a response time of a second or less on all tasks). The Dell solution

achieved 17 simultaneous virtual machines (VMs), while the HP solution delivered only 14 such VMs. Thus, the

Dell solution had a 21.4 percent performance advantage over the HP solution. (The Workload section explains in

more detail the virtualized

Exchange 2007 workload.)

Figure 2 illustrates the

performance/watt for each of

the two solutions. In this chart,

we normalized the results for

each workload to those of the

solution with lower

performance/watt. We

compute performance/watt by

dividing the benchmark’s

score by the average power

consumption, in watts, of the

solution during the period the

benchmark was delivering

peak performance.

As Figure 2 shows, the Dell

solution delivered 18.2 percent

more performance/watt than

the HP solution.

Principled Technologies, Inc., Virtualized Exchange workload performance comparison of end-to-end solutions with iSCSI storage

connections: Dell PowerEdge 2950 with Dell EqualLogic PS5000 vs. HP ProLiant DL385 with HP StorageWorks EVA 4400, October

2008.

Ibid.

Number of virtual machines

Virtualized Exchange 2007 workload results

HP ProLiant DL385

server and HP

StorageWorks EVA

4400 storage array

Dell PowerEdge 2950

server and Dell

EqualLogic PS5000

storage array

Figure 1: Virtualized Exchange 2007 workload performance results for the two server-and-

storage solutions we tested.

Summary of content (14 pages)

PAGE 1
TEST REPORT OCTOBER 2008 Virtualized Exchange workload performance comparison of end-to-end solutions with iSCSI storage connections: Dell PowerEdge 2950 with Dell EqualLogic PS5000 vs. HP ProLiant DL385 with HP StorageWorks EVA 4400 Executive summary Dell Inc.
PAGE 2
Performance/watt results Normalized comparison 1.40 1.20 HP ProLiant DL385 server and HP StorageWorks EVA 4400 storage array 1.00 0.80 0.60 Dell PowerEdge 2950 server and Dell EqualLogic PS5000 storage array 0.40 0.20 0.00 - Figure 2: Performance/watt results for the servers normalized to that of the HP ProLiant DL385 server and HP StorageWorks EVA 4400 storage array solution. Higher numbers are better.
PAGE 3
Server Dell PowerEdge 2950 server and Dell EqualLogic PS5000 storage array HP ProLiant DL385 server and HP StorageWorks EVA 4400 storage array Total number of virtual machines 17 14 Figure 3: The total number of virtual machines for the two server-and-storage solutions. Higher numbers are better. Figure 4 shows the average response time from the median run for each virtual machine.
PAGE 4
Test methodology Each server arrived with ESX 3i hypervisor installed on an internal USB key. Booting to VMware ESX Server 3i 3.5.0 We performed the following steps in order to boot to VMware ESX Server 3i 3.5.0 build 82664 on the server under test: 1. Power on the server. 2. Enter the Bios. 3. Enable the internal USB port. 4. Place the internal USB port at the top of the boot order. 5. Save and restart the server.
PAGE 5
7. 8. 9. 10. 11. 12. Click Close to close the Properties window. On the Networking configuration tab, click Add Networking. Select Virtual Machine for the connection type, and click Next. Under Create a virtual switch, select vmnic1, and click Next. Change the Network Label to Mail Network, and click Next. At the summary page, click Finish. Creating a virtual machine The steps below describe how we created a base virtual machine.
PAGE 6
19. 20. 21. 22. 23. 24. 25. Click Apply. Click Change power settings. Click Change plan settings under Balanced. Change the Turn off monitor drop-down menu to Never. Click Save changes and close the Power Options, Screen Saver Settings, and Personalization windows. Open the Start menu, type regedit in the search field and press Enter.
PAGE 7
o. At the Completing the Active Directory Installation Wizard screen, click Finish. p. When the system prompts you to reboot the server, click Restart Now. q. After the server reboots, log in to the system. Installing Internet Information Services 6.0 1. Select StartÆAdministrative ToolsÆServer Manager. 2. Click Add Roles. 3. At the Select Server Roles screen click Web Server (IIS). 4. In the pop up Add features required for Web Server (IIS) window, click Add Required Features. 5. Click Next. 6.
PAGE 8
14. In the Completion page, click Finish to close the New Public Folder Database window. 15. Close the Exchange Management Console, and reboot the server. 16. Once the server successfully reboots, log in, and select StartÆAll ProgramsÆAdministrative ToolsÆActive Directory Users and Computers. 17. In the Active Directory Users and Computers window, select test.com in the left pane, and open Users in the right pane. 18. Right-click a blank area in the Users folder, and select NewÆUser. 19.
PAGE 9
63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. Click Customize next to Maintenance interval. Remove all blue from the boxes so the system will not perform maintenance, and click OK. At the Mailbox Store 2 Properties window, click Apply. When a warning message appears, click Yes. When the message that the database files have been moved appears, click OK. Click OK to close the Mailbox Store 2 Properties window.
PAGE 10
3. 4. 5. 6. 7. 8. Assign a computer name of “Clientx” for the mail client, where x is the client number. For the licensing mode, use the default setting of five concurrent connections. Enter a password for the administrator log on. Select Eastern Time Zone. Use typical settings for the Network installation. Use “Testbed” for the workgroup. To set up this server, we had to install several additional software components. The following subsections detail the necessary installation processes.
PAGE 11
15. Once installation is complete, click Finish. 16. Download and install Load Generator using all defaults. Preparing Load Generator 1. Log into the MailClient using the tester account. 2. Select StartÆAll ProgramsÆMicrosoft ExchangeÆExchange Load Generator. 3. When the Load Generator window appears, select Start a new test. 4. Select Create a new test configuration, and click Continue. 5. Enter tester as the account name. 6.
PAGE 12
Appendix A – Server configuration information This appendix provides detailed configuration information about the servers, which we present in alphabetical order.
PAGE 13
Server Speed in the system currently running @ (MHz) Timing/Latency (tCL-tRCD-iRPtRASmin) Size Number of RAM modules Chip organization Hard disk Vendor and model number Number of disks in system Size Buffer size RPM Type Controller Controller driver Operating system Name Build number File system Kernel Language Graphics Vendor and model number Chipset Type Memory size Resolution Network card/subsystem Vendor and model number Type Driver Optical drive Vendor and model number USB ports Number Type Dell Power
PAGE 14
Appendix B – Storage array configuration information This appendix provides detailed configuration information about the array systems, which we present in alphabetical order. Enclosure General dimension information Height (inches) Width (inches) Depth (inches) U size in server rack Power supplies Total number Wattage of each Cooling fans Total number Dimensions (H x W) of each Voltage Amps Hard disk Dell Equal Logic PS5000XV HP EVA 4400 5.5 19.0 21.5 3 3.5 19.0 22.0 2 2 450 2 575 4 4x4 12 volts 1.