As awareness grows of the possibilities for data warehousing, many data warehouses for large projects are retrenching to make sure they are housed on platforms that can scale to meet their future needs. As part of this assessment, the option of using vendor benchmark facilities is often considered.
Many projects that need this service do not utilize it. The reverse is also true. Many use it for projects that do not have unique characteristics that market research would be just as useful for. If your project truly has special needs, consider benchmarking your performance at a vendor facility.
Since there usually is no direct cost for serious buyers, the major downside is the staff time that will be dedicated to the effort. This includes time to educate the vendor, compile the tests, collect representative data, possibly scrub it and finally run the tests and interpret the results. And if multiple architectures are involved, you're also trying to keep it apples-to-apples and keep it "fair" as much as possible despite the varied directions that the vendors would take the effort.
Be sure to compile true representative queries, test load times and, if applicable, cube build times. And be sure to "play" -- going well beyond the bounds of your pre-built queries and beyond what the vendor knows they are tuning the environment for! The data warehouse is supposed to be interactive and accommodating to all manner of "what if" and "on the fly"
Requires Free Membership to View
When you register, you'll begin receiving targeted emails from my team of award-winning editorial writers on the latest customer relationship management (CRM)and call center technology issues today. Our goal is to keep you informed on the hottest issues facing this fast-changing industry.
Hannah Smalltree, Editorial DirectorFinally, note the effort required on the part of the vendor to pull this together. Keep in mind you will have to pick up the ongoing work -- or buy it from the vendor. Note the data explosion from its raw form and note the amount of indexing, summary tables and other creativity required to accomplish the performance you see. A more manageable, predictable environment should sometimes trump a slightly better performing one.
For more information, check out SearchCRM's Best Web Links on Data Warehousing.
This was first published in July 2002
Join the conversationComment
Share
Comments
Results
Contribute to the conversation