My performance test environment was changed from RAID-5 to RAID-1. So I took the opertunity to run several before and after benchmarks so that the diffierence could be qualified.
The reality is there was none. The response time from LoadRunner didn’t change and even the read response time recorded by Oracle Stats Pack didn’t change as can be seen in the graph below:
I spoke to the storage manager about this and got the following reply “in most modern storage controllers, all I/O is satisfied from memory. You never get your data directly from disk and you never write directly to disk either. As such the theoretical differences in performance across the myriad of RAID implementations are masked by the array’s internal systems: notably cache and the sophisticated algorithms that deliver pre-fetching and other advanced capabilities.…”