Allocation and Migration of Virtual Machine using ABC and SVM based Optimization Strategy

Abstract

Cloud computing is one of the widely used technology to handle massive data generated at the cloud data center. To analysis this big data number of VMs are generated at the server end to balance load of PMs. When VMs are allocated to PMs resources are utilized to power this system. This process at times involves wastage of resources such as energy and Service Level Agreement (SLA) that drastically increases when unnecessary migrations occur to balance the overloaded PMs. Therefore, in the paper author proposes an energy efficient VM allocation and migration framework that is wise enough to minimize the number of migrations and reduces the instances of SLA violations. The idea here is to rank the available VMs according to their current load while considering the dynamics of energy and SLA violations using hybrid optimization involving Artificial Bee Colony (ABC) and Support Vector Machine (SVM) followed by machine learning architecture. The simulation analysis using 1000 iterations demonstrated reduced energy consumption and SLA violations achieved by minimizing the required number of migrations for load balancing of the PMs. This proved to achieve an effective resource management of cloud data centers.