Procurement of linux server config for obiee based on user count

User_DBCL0

HI,

Suppose we have 500 user going to user obiee12c dashboards. then what is the configuration we should have.

I am trying to get the information like how do come to conclusion on obiee hw config based on user count.

Is there any document/information please share here.

Gianni Ceresa

Let me answer you with a question: when are your 500 users going to be on your obiee? And what are they going to do exactly?

Because 500 concurrent users running heavy dashboard (containing many heavy analysis) is different than 500 users in total with a maximum of 5 concurrent users looking at dashboard consisting in a single table with 10 cells in total ...

Define a bit more the concurrent number of user (that's the one that really matter) and the kind of work they are going to do (consume prebuilt content, data pump via dashboards, power users etc.).

User_DBCL0

Thanks for your reply Gianni Ceresa,

Yes you are correct.

I have 50 concurrent users accessing 10 dashboards having 4 medium reports on each dashboard.

They will view the reports/download in to excel.

Please let me know if you need any more details

[Deleted User]

684461 wrote:Thanks for your reply Gianni Ceresa,Yes you are correct.I have 50 concurrent users accessing 10 dashboards having 4 medium reports on each dashboard.They will view the reports/download in to excel.Please let me know if you need any more details

"medium dashboards" is vague as well and "thy will download" likewise.

How much data gets pulled in? A good config where a couple of 100s of rows gets process? Vs a really bad ones where millions of rows are forced into the frontend and processed there?

It's very very hard to give you an answer on this one since I can work with literally gigabytes worth of data in multi-million row facts on my laptop in subsecond performance just like I can destroy a 12 core, 64GB RAM server with some really badly build things on much smaller data sets.

User_DBCL0

Yup. Thanks.

I will check the things at my end about the usage of dashboards