Performance Optimization - Professional SQL Server Analysis Services 2005 with MDX - page 627

Figure 13-2
Note that while we expect to see the general shape of the curves described
above, the limits will be different for your hardware, your cube design, the
queries your users execute, and the frequency with which queries are ex-
ecuted. With respect to Analysis Services, in a typical query scenario, CPU,
memory, and disk read speed can all be potential bottlenecks for your system.
In a cube or dimension processing scenario, CPU, memory, disk writing
speed, and network speed between the relational source and Analysis Ser-
vices can all be candidate areas for system optimization work.
For Analysis Services performance we refer to four main areas — database
design and processing, querying, and configuring Analysis Services (setting
properties). The first involves design optimization to facilitate the optimized
processing of dimensions and cubes. The second relates to optimization to
speed up MDX queries. The third involves fine tuning Analysis Services or
appropriate hardware acquisition based on requirements. Does performance
imply both scalability and optimization of Analysis Services? Depending on
whom you ask, one or both are true. The bottom line is that you need best
query performance with regard to your OLAP cubes. That is true regardless
of size, and however you can get it that doesn't involve the violation of federal
or state laws is fine. In the next section you will learn some of the design
techniques that can help in improving your Analysis Services performance.