An investigation on energy and deadline aware scheduler for optimization in virtual clouds

Abstract

The rapid growth of cloud computing has come along with considerable challenges in regards to task scheduling and resource allocation compelled by the unpredictable and dynamic nature of user demands. In cloud environments, different computing resources like CPUs, memory, bandwidth, and storage are shared by virtual machines. Specifically, resource allocation efficiency is crucial, as unexpected variations in task demands can result in under- or overutilization, leading to increased energy consumption, prolonged execution times, and higher operating costs at the expense of Service Quality assurances to the final consumers. newlineThis research suggests a method for multi-objective optimization to improve efficiency of cloud computing by carrying out task scheduling and resource allocation. The proposed model will enhance the performance of cloud infrastructures by optimizing throughput, reducing energy consumption, and decreasing execution time. These challenges will be addressed using meta-heuristic algorithms featuring better guarantee ratios, improvement in resource utilization, maximization of energy savings, among others. newlineThe first contribution involves the proposal for the improvement of the traditional Artificial Bee Colony (ABC) algorithm. Conventional ABC is good for exploration at the same time suffers from premature convergence and, hence, lacks any guarantee of optimality. In order to solve this, the Ogive Folded Cumulative Distribution function is presented, which encourages further exploration of interesting areas in the search space and produces more precise solution outcomes. newline