Benchmarking File System Benchmarking: It *IS* Rocket Science

Abstract

The quality of file system benchmarking has not improved in over a decade of intense research spanning hundreds of publications. Researchers repeatedly use a wide range of poorly designed benchmarks, and in most cases, develop their own ad-hoc benchmarks. Our community lacks a definition of what we want to benchmark in a file system. We propose several dimensions of file system benchmarking and review the wide range of tools and techniques in widespread use. We experimentally show that even the simplest of benchmarks can be fragile, producing performance results spanning orders of magnitude. It is our hope that this paper will spur serious debate in our community, leading to action that can improve how we evaluate our file and storage systems.