Q: We are in the process of comparing benchmarks for various enterprise-wide data warehousing platforms. How do I define the performance benchmarks? What should be the definition of complex, medium and simple query, and what should be the expected response timings?

Sid Adelman's Answer:

The vendors who run benchmarks will be happy to give you their profiles of simple, medium and complex but becauseyou are planning to implement something in your environment, it's more important for you to use your own profiles especially for the complex queries. Validate your profiles against those of the vendors, but ask your power users about their complex queries, what they have experienced in the number of tables joined, the number of rows returned, and any machine-intensive calculations performed. You need to know the volumes and distribution among the simple, medium and complex. Because you are performing the benchmarks to determine your platform, and because you do not want to have to change out platforms as you grow, be sure to build the benchmark using volumes that are at least three times your expectations. With success, your volumes of data and your volumes of concurrent users will grow by three times.

Expected response times - I'd wait and see what the benchmarks demonstrate but do not let the benchmark numbers create expectations of what you will actually get in your environment. The most important thing to know about expected response times is that response times will be highly variable.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access