I’ve found myself getting worked up lately about some of the file system benchmark information and tests that have been published. Vendors use these benchmarks to tout how well their file system did compared to brand X, but file system benchmarks can’t be looked at in the same way that say, SPEC benchmarks can be looked at, as SPEC is designed so that vendors can’t really skew the results.
The big problem with file system benchmarks is that they are just the opposite of SPEC benchmarks; there are no standard file system benchmarks that are run and agreed upon by vendors. I was a benchmarker at one time and consider myself a reformed benchmarker, so I know the tricks that vendors can use. We’ll look at some of them and hopefully in the process make you a more informed consumer…
