In the storage world, performance is a big deal and is often a differentiator….how do you ask? Well, some users will make a decision on Fibre Channel vs. iSCSI based on throughput performance, some users will make a specific purchasing decision on a device based on NFS performance published by SPEC SFS (see NFS performance for Windows Storage Server here)….some end users will make a decision based on IOmeter data. While these types of results can give a glimpse of how a device will perform in the production environment, it is still a glimpse only. For example, in order to get good benchmark results, vendors can artificially submit solutions with way more disk drives than what would be used in the production environment…vendors have the freedom to carve the storage infrastructure as they need to get good results. All of this is fine but end users have to realize that benchmark results can be far from what will be achieved in the production environment.
There are some benchmarks (one example is the Storage Performance Council benchmarks) and tools that can do a better approximation of the production workload. In the case of finding out about performance when running Exchange, a great resource to check is the Exchange Solution Review Program (ESRP). Net, net performance matters but make sure you know what you are getting from the benchmark results.