HP proliant dl585 g5 ユーザーズマニュアル

Test scenarios 

For the Heavy User type, HP initiated testing by running the appropriate script with a group of 15 

simulated users. Start times were staggered to eliminate authentication overhead. After these sessions 

finished, HP added 15 more users, then repeated the testing. Further groups of 15 users were added 
until the maximum number of users was reached. 
For Medium and Light User types, HP utilized groups of ten users. 

Performance and scalability metrics 
While the scripts were running, HP monitored a range of Windows Performance Monitor (Perfmon) 

counters to help characterize server performance and scalability. In particular, HP has monitored CPU 

utilization (% Processor Time) to establish the optimal number of users supported on an HP SBC server 

– by definition, the number of users active when processor utilization reaches 80%. At this time, a 
limited number of additional users or services can be supported; however, user response times may 

become unacceptable.  
To validate scalability metrics obtained using Perfmon, HP also runs canary scripts to characterize 

Heavy User response times – a very practical metric – for discrete activities such as an application 
being invoked or a modal box appearing. By monitoring response times as more and more users log 

on, HP has been able to demonstrate that these times are acceptable when the optimal number of 

users (as determined using Perfmon counter values) is active. 

Note: 

When running canary scripts, HP considers user response times to become 
unacceptable when they increase markedly over a baseline measurement. 

7