Is there any best practice to determine required resources for a typical BI ecosystem?

Iliaa

To be more clear, we have a data warehouse with approximate size of 10 Terabyte.

Our system should be able to provide fair (or excellent) responses to the requests of 10 users simultaneously.

I'm sure that the required resources (RAM, CPU, Hard Disk, type of OS, ...) are strongly depended on the types of Analyses being run,

but I want to know that, according to experience, what is a good primitive configuration for such a system?

Thanks in advanced.

P.S. I found this article but it seems that the article is tightly coupled to IBM machines.

rmoff

What kind of workload are you expecting? Pre-canned reports, ad-hoc, or both? Is your 10TB raw data, or does that include aggregates?

Iliaa

I'm new to OBIEE and I don't know what does mean Pre-canned reports and ad-hoc in this context.

10 TB is the size of our data warehouse, i.e. our Star Model (Dimensions and Facts), not Database raw data.

Manoj Dixit

Pre-canned report = a report built for a user or a group of users upon request by IT department.

Ad-hoc report = a report built by a user which was not planned or requested for before.

For example, a report is pre-built to display production quantity by day and hour. An end user notices that during 12 noon and 13, production is almost nil. The end user builds a report for the hour and reason and realises that the reason is lunch time. This is an ad-hoc report.