I've often wondered if the PeopleSoft community can learn anything about optimising or validating our PeopleSoft infrastructure environments by gathering
PeopleSoft Ping data from other users. So, I figured let's gather some data and start a discussion. I've created a
Google Form to gather some basic and anonymous data. The data gathered
is anonymous and will only be used for the purposes of community discussion around optimising infrastructure.
Please post questions, ideas and general infrastructure performance and optimisation comments against this blog post.
8 comments:
Graham - Great initiative!
Would it also be useful to include additional details such as Application (HCM, FSCM, CS, etc), Version (9.1, 9.2, etc), infrastructure details (number of app servers/domains, number of web servers/domains)?
Thanks for your ideas Sasank. The question of optimal configuration of PeopleSoft infrastructure is a wide and complex subject. With this blog post I wanted to just focus on something really simple like PSPing. Gathering App details and version numbers and numbers of servers/ domains shouldn't have an impact on PSPing performance - unless of course someone wants to suggest and prove otherwise.
Graham
Single core CPU performance on the PeopleSoft application server plays a key role in getting transaction times down especially transactions that are heavy on PeopleCode and light on SQL.
Core performance is key because a single application server transaction will consume as much of 1 and only 1 core as it can. So, the faster that core is the faster the transaction will run. Adding more cores will not make individual transaction times run faster. Adding more cores may help with high user throughput.
This is a useful tool for comparing CPU performance. http://cpuboss.com/compare-cpus
for example... http://cpuboss.com/cpus/Intel-Xeon-X5570-vs-Intel-Xeon-E3-1230
what could you concluded if you compare 2 infrastructures using the PSPING tool ?
@Stéphane The results are accessible using the "view spreadsheet" link on the main post here
>>what could you concluded if you compare 2 infrastructures using the PSPING tool ?
Great question. Here's where I'm coming from. If my Ping result for my web server is 0.9 and someone else reports one as low as 0.009 then I'd like to have some kind of discussion around how to make mine faster. With so many layers of technology between the user and PeopleSoft data it becomes a complex matter to do comparisons. However, there may be some obvious things that contribute to improving performance and PSPing may or may not tease that out.
Hi Graham,
To improve this topic, I think you shoud add a column with only Web + App + Db (and not network).
This would help to understand which scenarion is the best out of the list ?
Regards,
Stéphane.
Recommendations should also be mentioned, like if you note the high time in webserver time or browser time then what actions need to be performed.
Thanks
Post a Comment