StopGap

In order to reduce the energy consumption, virtualized infrastructure providers (IaaS) usually pack virtual machines (VMs) on as few servers as possible. This process, called VM consolidation, can be ineffective when the amount of available resources (hereinafter "holes") on a server is not enough for packing a new VM. These holes remain unused and therefore wasted. In order to avoid as much as possible the resource waste, we introduce StopGap which dynamically divides a VM into smaller "pieces" so that each piece fits into the available holes on the servers. However, the traditional cloud management policy was not conceived for these elastic VMs. Thus, we propose HRNM: a new resource allocation and negotiation policy in the IaaS. We have demonstrated the HRNM applicability by implementing a prototype compliant with two mainstream IaaS managers: OpenStack and OpenNebula. Finally, using Google data center traces, we show an improvement of about 62.5+ for the traditional consolidation engines.