![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
I know where to go to find single disk benchmarks, but is it accurate to say that comprehensive benchmarking of RAID controllers doesn't really exist out on the net, at least in a central repository using a formal methodology?
The metrics I'm thinking of,
- sustained throughput ceiling
- sustained IOPS ceiling
- array build and rebuild times
- resilvering of a hot spare
- CPU utilization under load
Right now, the best I can think of is to look at the IO engine (IOP, Highpoint XOR, ICHR, etc) and make guesses based on forum posts out on the net. It sure would be nice to establish a standard benchmark methodology, perhaps one that's at least partially resilient to spindle types since that will vary from benchmark to benchmark. As benchmarking goes, this seems like a relatively immature field in need of some structure.
The metrics I'm thinking of,
- sustained throughput ceiling
- sustained IOPS ceiling
- array build and rebuild times
- resilvering of a hot spare
- CPU utilization under load
Right now, the best I can think of is to look at the IO engine (IOP, Highpoint XOR, ICHR, etc) and make guesses based on forum posts out on the net. It sure would be nice to establish a standard benchmark methodology, perhaps one that's at least partially resilient to spindle types since that will vary from benchmark to benchmark. As benchmarking goes, this seems like a relatively immature field in need of some structure.