De-Fragmenting the Cloud

Abstract

Existing Virtual Machine (VM) placement schemes have looked to conserve either CPU and Memory on the physical machine (PM) OR network resources (bandwidth) but not both. However, real applications use all resource types to varying degrees. The result of applying existing placement schemes to VMs running real applications is a fragmented data center where resources along one dimension become unusable even though they are available because of the unavailability of resources along other dimensions. An example of this is unusable CPU because of a bottlenecked network link from the PM which has available CPU. To date, evaluations of the efficacy of VM placement schemes has not recognized this and it's ill effects, let alone try to measure it and avoid it. In this paper, we first define the notion of what we term relative resource fragmentation and illustrate how it can be measured in a data center. The metric we put forth for capturing the degree of is comprehensive and includes all key data center resource types. We then propose a VM placement scheme that minimizes this and therefore maximizes the utility of data center resources. Results of empirical evaluations of our placement scheme compared to existing placement schemes show a reduction of by as much as 15% and an increase in the number of successfully placed applications by as much as 20%.