My Account 
Measuring Storage Health
A large number of individuals tend to use server benchmarks in order to measure storage performance, and it seems to have caught on to the point where it has become a de facto standard. Not to say that server benchmarks aren’t without its uses, but it’s definitely not meant for measuring storage health, because it only measures the maximum performance of the server.
Additionally, in many of the cases where a server benchmark is used to measure storage performance, the server and OS used are not tuned or optimized, which meant that the storage itself isn’t being given the proper full load that it can handle. Ultimately, this makes the test worthless for measuring storage health because the true capabilities of the storage being tested are left on the table.
In order to truly measure storage performance, the backend storage must be the sole target resource being measured, avoiding processor bottleneck. Being cpu-bottlenecked is a common case because storage arrays have the capacity to handle several heavy workloads from multiple servers at the same time. There is a need to use storage benchmarks designed by storage engineers themselves.
True storage benchmarking tools measure storage health by giving a storage array a simulation of multiple workloads with blended patterns designed to reflect extremely heavy real world use. More complex storage benchmarks even use simulation of moving hot spots, skew, caching effects, and non-uniform access in order to truly provide the most accurate metrics related to storage health.
At the end of the day, it is important to use the right tool for the job. Server benchmarks – even the best ones – won’t do the job if what you need is measure the storage health. You need to use a storage benchmark that was specifically designed for measuring the performance of storage.
