Back close

An Efficient Proactive VM Consolidation Technique with Improved LSTM Network in a Cloud Environment

Publication Type : Journal Article

Publisher : Springer

Source : Computing

Url :

Campus : Bengaluru

School : School of Computing

Year : 2023

Abstract : The power consumption of datacenters is multiplying, and several survey reports stated that the power consumption of datacenters will reach approximately 8000 TWh by 2030 if do not utilize cloud resources effectively. To use allocated cloud resources effectively, one of the prominent solutions is the VM consolidation technique. VM consolidation technique manages cloud resources effectively while simultaneously satisfying the objectives of cloud users and providers. Additionally, it helps to increase servers’ performance while reducing the high power consumption of datacenters. However, unnecessary actions of VM consolidation technique cause unsuitable VM selection and inappropriate VM placement, which degrades resource management performance, poor QoS, and SLA violations. To overcome this issue, this paper proposed a resource, SLA, power-aware proactive VM consolidation technique by using an improved LSTM network to manage the allocated resources effectively. The proposed proactive VM consolidation technique helps reduce the high power consumption of datacenters while maximizing resource management performance and avoiding SLA violations. Finally, the authors measure the proposed methodology effectiveness by considering the benchmark dataset of NASA servers, and experimental results proved that an improved LSTM network can able to achieve an average accuracy rate of up to 94% with minimum prediction error rate. Proactive VM consolidation technique minimized nearly 30% of the power consumption of datacenters compared with conventional VM consolidation technique.

Cite this Research Publication : K Dinesh Kumar and E Umamaheswari, “An Efficient Proactive VM Consolidation Technique with Improved LSTM Network in a Cloud Environment”, Computing, Springer. ISSN: 1436-5057.

Admissions Apply Now