HP ProLiant 4500 Performance Analysis and Capacity Planning for Microsoft Site - Page 13

PS User Scenario

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 13 highlights

WHITE PAPER (cont.) 1...3 Operating System ILS Software Disk Subsystem Not enough information Not enough Information No Unless ILS implementation details are known, you cannot determine whether performance bottlenecks are in the operating system or ILS. Unless ILS implementation details are known, you cannot determine whether performance bottlenecks are in the operating system or ILS. There is almost no paging and other disk accesses are low. Personalization System (PS) Performance Measurement Procedures Testing PS requires a tool that uses the standard HyperText Transfer Protocol (HTTP) of the Web because PS is accessed via ASP-enabled Web pages served by IIS. With this in mind, both WCAT and perfmon were used for the PS performance measurements. WCAT was selected because it can create a reliable, reproducible SUT load using few clients, and because it summarizes the performance statistics gathered from all of the client systems. While InetLoad seems better suited to answer the question "How many users can the SUT support," the current release, Release 2.0, is sufficiently weak in reporting and unstable enough to merit the use of WCAT at this time. The performance of an SUT will depend on the size of the user community that it serves because PS keeps a database of user information for personalizing Web pages. A User Property Database (UPD) was created on the SUT to represent the size of the user community of interest. This was done using an ASP-enabled Web page and a Web browser to initialize a UPD for the number of users desired. UPDs of various sizes were also created. Which one was used for testing was managed by renaming them to the name specified via the User Property Database Administration Tool, which is provided as part of the PS. The PS performance tests were run on an SUT for user communities of various sizes. PS User Scenario With the goal of determining how many users the SUT can support, a scenario representing a typical user at a 100% personalized Web site was developed (see Table 12). That is, each Web page was delivered to the user via an ASP that accessed the UPD and customized the Web page accordingly. Such a Web site represents a worst-case scenario for measuring PS performance, since a Web site with less personalization will be able to support higher transaction rates because it does not access as many ASP-enabled Web pages.

  • 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.)
13
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Operating
System
Not enough
information
Unless ILS implementation details are known, you cannot
determine whether performance bottlenecks are in the
operating system or ILS.
ILS Software
Not enough
Information
Unless ILS implementation details are known, you cannot
determine whether performance bottlenecks are in the
operating system or ILS.
Disk
Subsystem
No
There is almost no paging and other disk accesses are low.
Personalization System (PS) Performance Measurement Procedures
Testing PS requires a tool that uses the standard HyperText Transfer Protocol (HTTP) of the Web
because PS is accessed via ASP-enabled Web pages served by IIS. With this in mind, both WCAT
and
perfmon
were used for the PS performance measurements. WCAT was selected because it can
create a reliable, reproducible SUT load using few clients, and because it summarizes the
performance statistics gathered from all of the client systems. While InetLoad seems better suited
to answer the question “How many users can the SUT support,” the current release, Release 2.0,
is sufficiently weak in reporting and unstable enough to merit the use of WCAT at this time.
The performance of an SUT will depend on the size of the user community that it serves because
PS keeps a database of user information for personalizing Web pages. A User Property Database
(UPD) was created on the SUT to represent the size of the user community of interest. This was
done using an ASP-enabled Web page and a Web browser to initialize a UPD for the number of
users desired. UPDs of various sizes were also created. Which one was used for testing was
managed by renaming them to the name specified via the User Property Database Administration
Tool, which is provided as part of the PS. The PS performance tests were run on an SUT for user
communities of various sizes.
PS User Scenario
With the goal of determining how many users the SUT can support, a scenario representing a
typical user at a 100% personalized Web site was developed (see Table 12). That is, each Web
page was delivered to the user via an ASP that accessed the UPD and customized the Web page
accordingly. Such a Web site represents a worst-case scenario for measuring PS performance,
since a Web site with less personalization will be able to support higher transaction rates because
it does not access as many ASP-enabled Web pages.