Chapter 7. Maximizing Mondrian performance
This chapter is recommended for
| Business analysts | |
 |
Data architects |
|
Enterprise architects |
| Application developers |
Adventure Works analysts have been generally happy with the Mondrian’s abilities. They like the reports and dashboards and particularly being able to do analysis on the fly. Some have even become proficient with MDX queries for performing advanced analysis. But as the amount of data grows, some of the reports and analyses are starting to feel sluggish, and not as quick as users demand.
One of the promises of Mondrian is that it supports analytics at the speed of thought. This means that when an analyst makes changes to a report, such as adding or removing dimensions and measures, adding calculations, and applying filters, the report needs to be updated within seconds, rather than minutes or hours. Given that analysis is often done over millions of records, performance is extremely important.
Out of the box, with a well-designed star schema, Mondrian performance is very good for a wide variety of datasets and queries. But some businesses want to do real-time analysis against millions of facts and thousands of dimension members. Eventually, even the fastest database and software will start to bog down with straight database calls. By default, Mondrian will perform some caching to speed things up, but squeezing out the highest levels of performance from your data sometimes takes additional configuration and effort.