The "Big Iron Test"

Posted by Benny Tritsch on November 25, 2004 – updated on January 31, 2005

[Introduction] [Test Setup] [Testing Methodology] [Analysis 1] [Analysis 2] [Analysis 3]

Testing Methodology

For the first two tests Windows Server 2003, Enterprise Edition (one of the Windows Server 2003 family of operating systems) was installed. For the third test, Windows 2000 Server, Enterprise Edition was installed. For both operating systems Office 2000 and Adobe Acrobat Reader 6 were installed.

For each test scenario, the test engineer started groups of ten client sessions on the client machines with an interval of some seconds between each session. To do so, the test engineer needed to create the session by running the script Cr.cmd adding an individual number between 1 and 150 as an attribute for each user. After the initialization of a session the logon script checked if the documents required for the test scenario were available on the desktop of the user. If not, they were copied there from a standard source. Finally, every ten seconds one of the applications with a corresponding document was started.

• Notepad with Terminalservices.txt
• Adobe Acrobat Reader 6 with PAG3.pdf
• Microsoft Word 2000 with FSCDemo.doc

After the group of ten client sessions was started, a short stabilization period was observed in which no additional sessions were started. After the stabilization period, the test engineer started the next user sessions on another client machine; this prevented any interference between each group of ten client sessions.

Each user session showed a memory footprint of approx. 60 megabytes when all applications were launched.

During the tests the following performance counters were monitored and analyzed:

• Memory: Available MBytes
• Memory: Free System Page Table Entries
• Memory: Page Faults/sec
• Memory: Pages/sec
• Memory: Pool Nonpaged Bytes
• Memory: Pool Paged Bytes
• Physical Disk: Avg. Disk Queue Length
• Processor: % Interrupt Time
• Processor: % Processor Time
• Processor: Interrupts/sec
• Server: Pool Nonpaged Bytes
• Server: Pool Paged Bytes
• System: Context Switches/sec
• System: Processes
• System: Processor Queue Length
• Terminal Services: Active Sessions

The following chapters present the analysis of some of the performance counter values observed.

Next