HP ProLiant 4500 Performance Analysis and Capacity Planning for Microsoft Site - Page 14

PS Test Procedures, PS Test Results

Get HP ProLiant 4500 PDF manuals and user guides View all HP ProLiant 4500 manuals
Add to My Manuals
Save this manual to your list of manuals

Page 14 highlights

WHITE PAPER (cont.) 1...4 Table 12: PS User Scenario Operation # Operation 1 Create a new UPD entry 2 Get a personalized Web page 3 Get a personalized Web page and update personalization information Operation Frequency 5% 91% 4% Comment 5% of the users will be new, so create a new UPD entry for each of them. Each Web page will consist of an HTTP GET of the .asp file and 4 HTTP GETs of .gif files. Each .asp file accessed will be 8,290 bytes long and each .gif file will be 5,480 bytes long. This does the same HTTP GETs as does Operation #2 and also updates 2 properties in the UPD for the simulated user. As mentioned, other user scenarios are possible and may lead to varied platform transaction capacities. An Excel spreadsheet has been provided in Appendix A to address this issue. PS Test Procedures Before starting any tests, UPDs were created to simulate user communities of 1,000, 10,000, 100,000, and 1,000,000 people. Then, the following basic PS test procedures were used: • At SUT boot time, automatically start IIS. IIS logging is enabled. • Rename the UPD with the appropriate-sized user community to the active one used by PS. • Run perfmon on the SUT and have it log the processor, Active Server Pages, HTTP Service, memory, and User Property Database performance data to a file, as well as show a real-time chart of the CPU utilization and the appropriate transaction rates. The Network Interface and Physical Disk may also need to be monitored, depending on configuration performance. • Start the tests running on the client systems. The WCAT collects performance statistics and logs them to a file. • At the end of each test, collect the perfmon log from the SUT, the PS log file, and the client performance data file. • Reboot the SUT and test for the next-sized user community. All tests were done on isolated LANs that were inactive except for the test traffic. Tests were run with a five minute warm-up period and a 10 minute measurement period using WCAT. PS Test Results Figure 3 summarizes the ProLiant 2500 PS performance measurements shown in Table 13. Similarly, Figure 4 summarizes PS performance for the ProLiant 7000 based on the data in Table 14.

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
W
HITE
P
APER
(cont.)
14
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Table 12: PS User Scenario
Operation #
Operation
Operation
Frequency
Comment
1
Create a new UPD entry
5%
5% of the users will be new, so
create a new UPD entry for
each of them.
2
Get a personalized Web page
91%
Each Web page will consist of
an HTTP GET of the .asp file
and 4 HTTP GETs of .gif files.
Each .asp file accessed will be
8,290 bytes long and each .gif
file will be 5,480 bytes long.
3
Get a personalized Web page
and update personalization
information
4%
This does the same HTTP
GETs as does Operation #2
and also updates 2 properties
in the UPD for the simulated
user.
As mentioned, other user scenarios are possible and may lead to varied platform transaction
capacities. An Excel spreadsheet has been provided in Appendix A to address this issue.
PS Test Procedures
Before starting any tests, UPDs were created to simulate user communities of 1,000, 10,000,
100,000, and 1,000,000 people. Then, the following basic PS test procedures were used:
•
At SUT boot time, automatically start IIS. IIS logging is enabled.
•
Rename the UPD with the appropriate-sized user community to the active one used by PS.
•
Run
perfmon
on the SUT and have it log the processor, Active Server Pages, HTTP Service,
memory, and User Property Database performance data to a file, as well as show a real-time
chart of the CPU utilization and the appropriate transaction rates. The Network Interface and
Physical Disk may also need to be monitored, depending on configuration performance.
•
Start the tests running on the client systems. The WCAT collects performance statistics and
logs them to a file.
•
At the end of each test, collect the
perfmon
log from the SUT, the PS log file, and the client
performance data file.
•
Reboot the SUT and test for the next-sized user community.
All tests were done on isolated LANs that were inactive except for the test traffic. Tests were run
with a five minute warm-up period and a 10 minute measurement period using WCAT.
PS Test Results
Figure 3 summarizes the ProLiant 2500 PS performance measurements shown in Table 13.
Similarly, Figure 4 summarizes PS performance for the ProLiant 7000 based on the data in Table
14.