Proficient Resource Allocation Technique for Cloud Resource Allocation using Deep Learning

Abstract

Cloud Provider (CP) offers resources to the various categories of clients according to the consumer's required demand for quality of service (QoS). When a physical machine (PM) is overloaded, the performance of its virtual machines (VMs) may degrade. Idle PMs can be shut down to conserve energy. This paper introduces a new approach for resource provisioning through VM consolidation and migration, aiming to meet user demands, minimize Service Level Agreement (SLA) violations, and reduce performance degradation during resource shortages. Initially, the workload of PMs for future time interval is predicted from the workloads of several previous time intervals of PMs using deep learning. If resource utilization across PMs is uneven, the resource provisioning method is regularly activated during these intervals.