Controlling Storage I/O Utilization

For vSphere, controlling memory or CPU allocation and utilization is relatively easy. The hypervisor determines how busy the CPU is and whether physical memory has been depleted. When resource contention occurs for either of these resources, not only is it easy to detect, but it is also easy to correct. If the CPU is too busy and there are not enough CPU cycles to go around, then you do not schedule cycles for lower-priority VMs—instead, you assign more cycles to higher-priority VMs.

How is this priority determined? Recall that Shares values are the mechanism that vSphere uses to determine the priority. Likewise, if memory becomes constrained, the balloon driver is inflated in the VMs' guest OS. It then reclaims some memory or slows down the rate of allocation to lower-priority VMs and increases the rate of allocation to higher-priority VMs. Not only does the hypervisor have complete visibility into the utilization of these resources, it...