Energy Aware Resource Allocation Based On Optimization and Minimum Migration Time

Abstract

Recently Cloud Computing is an emerging arena of research where virtual machine has an eminent role for affording service to the clients through computing resources all over the world on demand basis. Virtual machine (VM) availability awareness is one of the prominent tasks scheduling technology which is greatly utilized in cloud platform. The major challenge involved in VM availability is that dynamic change and uncertainty involved in task scheduling along with task services quality necessities which in turn extremely impacts on   cloud task scheduling capability.  The optimal scheduling currently utilizes Multi-level queue scheduling Particle Swarm Optimization algorithm (MQPSO) while it does not focus on host utility detection. Various failures such as VM failure, Connection failure, and Response failure may occur due to the greater exploitation of host which in turn causes greater energy consumption. Hence improved cuckoo search optimization is greatly involved for Multi-level queue scheduling through validating the combination of Shortest-Job-First (SJF) buffering in addition to Min-Min Best Fit (MMBF) scheduling algorithms (SJF-MMBF). Adaptive Neuro Fuzzy Inference System (ANFIS) algorithm is additionally utilized for over utilized host detection and Virtual Machines (VMs) migration from over-utilized hosts to the other hosts is accomplished and thereby mitigating energy consumption. The various assessment parameters, like Throughput, Delay, Cost, and Energy consumption are assessed for the projected load balanced task scheduling approach by means of Experimental analysis and compared with the prevailing approaches.