Enterprise Data Warehouse Benchmarks
Posted: 19 May 2011 | Source: Teradata
The historical and continuing high rate of IT project failure and budget and schedule overruns, coupled with senior management’s increasing insistence on quantifiable investment justification, leave IT in a position today of needing to control cost and mitigate risk like never before. Many organizations are facing the need to replace their data warehouse with a second- or even a third-generation implementation.
Many are considering an enterprise data warehouse (EDW) that goes beyond their earlier data marts – an EDW that incorporates all or most of the areas that run the business and from which critical strategic, tactical, and sometimes operational decisions are made. They have several options in terms of vendors, technology, and architecture – and they can’t afford to make a bad decision. One very effective way of ‘test-driving’ the options to see how well they’ll perform in one’s actual environment is through benchmarking. With first generation data warehouse implementations, many IT shops did not bother to use any benchmarking, instead opting to adapt their existing OLTP database for their datawarehouse. What they have learned is that not only are no two databases alike, but also that no two implementations of the same database are alike. Others settled for using industry-standard benchmarks as a guide instead of conducting their own benchmarks. If you’ve seen one benchmark, you’ve only seen one benchmark. ‘Standard benchmarks’ bear little resemblance to the real world.