Since performance is highly depended on many variables so there isn't a single answer to the question. There are two parts to the performance - 1) hardware and 2) the design of the cube and the reports.
For hardware, in general majority of the heavy processing happens on SQL Server Analysis Services (SSAS) to process all the queries from Analyzer so we always recommend to use a server class machine to host the SSAS with room to scale-up for both CPU and memory. For Analyzer server (IIS Server), although it doesn't have to do much of the processing but it does have to hold all the sessions from all the users to deliver the zero-foot-print experience therefore the key focus here is memory. For IIS/Analyzer server we recommend to use a 64bit machine starts with at least 4 to 8 GB of memory with room to expand for more. Other performance factors are network bandwidth and conditions that also plays a role in speedy delivery the data. We recommend to scale-up first before scale-out. To scale-out, you can always add additional load balance instance of Analyzer to join the main instance.
Analyzer Export Service (AES) is a new feature introduced in Analyzer version 3.0.2359 that renders report to Excel, PDF and other formats. This is potentially a process intensive process depending on the number, size and type of the reports, so if you plan to deliver reports via scheduling then you might need to take this into consideration. Luckily AES can be installed on the same server as Analyzer or on any other servers to share the process load.
From report performance, please see Report Performance and Best Practice.
0 Comments