An Energy Aware Cost Recovery Approach for Virtual Machine Migration

Abstract

Abstract. Datacenters provide an IT backbone for today's business and economy, and are the principal electricity consumers for Cloud computing. Various studies suggest that approximately 30% of the running servers in US datacenters are idle and the others are under-utilized, making it possible to save energy and money by using Virtual Machine (VM) consolidation to reduce the number of hosts in use. However, consolidation involves migrations that can be expensive in terms of energy consumption, and sometimes it will be more energy efficient not to consolidate. This paper investigates how migration decisions can be made such that the energy costs involved with the migration are recovered, as only when costs of migration have been recovered will energy start to be saved. We demonstrate through a number of experiments, using the Google workload traces for 12,583 hosts and 1,083,309 tasks, how different VM allocation heuristics, combined with different approaches to migration, will impact on energy efficiency. We suggest, using reasonable assumptions for datacenter setup, that a combination of energy-aware fill-up VM allocation and energy-aware migration, and migration only for relatively long running VMs, provides for optimal energy efficiency.

Energy efficient computing, clusters, grids and clouds: A taxonomy and survey

“…Migration would be more feasible if the VM is migrated to a more energy efficient target server where the VM will be able to recoup back its migration energy overhead. A comprehensive research is required to know which VMs (of which runtime) should be able to recoup back their migration energy and how much energy savings can be made if the VM is subsequently running more efficiently [45]. We found in a simulated experiment, that server consolidation is more energy efficient if the idle servers are switched off, as they are consuming more energy.…”

“…Server consolidation ensures that IT setup subsidizes as little as possible to CO 2 emissions, and recover cooling & power capacity which exponentially drop energy costs. However, we found in our previous work [45], that consolidation is some time more expensive and cost extra energy due to migrations. Largely, efficient scheduling (resource management) techniques are more cost-energy efficient than migration based methods.…”

“…As discussed earlier, migration would be more costly if the VM (in migration) got terminated during or just after the migration process is completed. A VM migration would be more economical and energy efficient if the migration is performed to a more energy efficient target server, enabling the VM to recover back its migration cost and runs subsequently more efficiently on the target server [45]. Table 12 summarizes different approaches (discussed in this survey), based on system characteristics (clusters, grids, clouds) and the approach (scheduling, virtualization, consolidation etc.)…”

“…Therefore new VM level energy consumption models are needed to measure accurate levels of total energy consumption across virtualized platform [44]. In our previous work [45], we proposed an energy consumption model for virtualized systems, to decide which type of host could run a VM more energy efficiently. Ibrahim at el.…”

“…In cluster and grids, a server efficiency factor like the one used by [112], might be of significant importance to run the task on most efficient server. However, in virtualized platform, as the efficiency of a server is dependent on how many VMs are running on it (the baseline idle power in divided amongst all running VMs), new metrics like the one proposed in [45] are sill required.…”