We us NFS as our storage protocol and have seen very poor performance on our VM’s from time to time. On our SAN, the average latency is around 2 – 4ms which is ok. What I have observed is when our host our pegged at 95 – 100% CPU usage, our DAVG time goes up to 15 – 25ms. With that being said is the following state correct. With file level storage when the host is CPU constrained it will cause delays in the NFS calls to the SAN which in turn causes poor performance. In short, the CPU usage on the host has a direct result on NFS performance. This is not the case with block storage though correct?