Adaptive Energy-Aware Algorithms to Minimize Power Consumption and SLA Violation in Cloud Computing

Abstract

Objective: With the establishment of virtualized datacenters on a large scale, cuttingedge technology requires more energy to deliver the services 24*7 hours. With this expansion and accumulation of information on a massive scale on data centers, the consumption of an excessive amount of power results in high operational costs. Therefore, there is an urgent need to make the environment more adaptive and dynamic, where the overutilization and underutilization of hosts are well known to the system and active measures can be taken accordingly. To serve this purpose, an energy-efficient method, for the detection of overloaded and under-loaded hosts, has been proposed in this paper. For implementing VM migration, VM placement decision has also been taken to save energy and reduce SLA (Service Level Agreement) rate over the cloud. Methods: In the paper, a novel adaptive heuristics approach has been presented that concerns with the utilization of resources for dynamic consolidation of VMs based on the mustered data from the usage of resources by VMs, while ensuring the high level of relevancy to the SLA. After identification of under-load and overload hosts, VM placement decision has been taken in the way that takes minimum energy consumption. A minimum migration policy has been adopted in the proposed methodology to minimize execution time. The validation of the effectiveness and efficiency of the suggested approach has been performed by using real-world workload traces in the CloudSim simulator. Results: The results shows that the proposed methodology is ideal for SLA, but costs more for VM migration. Conclusion: A deep analysis must be done in existing energy efficient approaches and a new platform should be suggested to save energy in real life.