System z: Accurate benchmarking tips

Thursday, December 25, 2014
Earlier this year, September 2014 to be exact, I wrote my first article for the online version on the IBM Systems Magazine.  The article was titled "An accurate benchmark is important".

In this article, I talk about some best practices for designing benchmarks that are meant to compare the same workload on various platforms.  Specifically, I call out reasons why benchmarks might "miss the mark" and will not yield realistic results that you can make sound business decisions on.

Surprisingly, designing a "real-world" benchmark that is entirely accurate is not as simple as running the same workload on multiple servers.  There is quite a bit more to it than that.....and having inaccurate information (no matter how scientifically you generate it) is still inaccurate.

Before you design your next benchmark, why don't you consider some of the suggestions in this article and run a well-designed benchmark?  Even if you don't agree with all of my points, at least you thought about it and made a choice that you can hang your career on.

My article, "An accurate benchmark is important", is available to read at http://www.ibmsystemsmag.com/mainframe/tipstechniques/systemsmanagement/accurate-benchmark/