Resource Optimization and Dynamic Workload Prediction Strategy for Large Scale Cloud Environment

Abstract

Cloud computing is a ubiquitous computing paradigm which eliminates the technical obstacles which the organizations have to deal with and provides IT services on-demand. Cloud computing is a cost-effective model for provisioning services and makes the ICT management effective in dealing with real-world workloads. It is becoming popular since it provides better usability at a lower cost leading to higher utilization and better management. Clustering based approach is used for the allocation of virtual machine to the incoming request. Here in the proposed system a hybrid approach is used with kernel based fuzzy C-means clustering mechanism with Bayesian prediction for the provisioning of resources in the heterogeneous cloud environment. The results are optimized with particle swarm optimization. Hence the total average waiting time is minimized and the results are evaluated using Cloudsim toolkit and the results reveals that the hybrid system achieves better performance when compared to the existing technology.