SNIA and hyperscalers are working together on storage standards to better optimize for their particular needs. SNIA’s Mark Carlson recently gave a talk describing how the DePop standard was developed to get around the problem of tail latency seen in these massive scale-out data centers.
Viktor van den Berg explains how vSphere storage latency should be determined, and why blocksize is an important matter: So, let’s say we have a latency that’s bigger then 30ms, do we have problem? Not necessarily, the SCSI block size also determines the latency value. The SCSI block size is depending of the type of […]