A Novel Approach for Cloud Computing Load Balancing with Machine Learning

Abstract

Load balancing deal with the task of uniform distribution of work load among the machines, no machine should exist with the heavy work load or less work load. Appropriate online solutions for LB depend on the computing environment and the jobs. Inspired by the present recent trends in computer technology, machine learning approach adopted to develop a scheduling strategy for load balancing. A Novel approach presented by the linear regression approach. By using this approach VM migration can be avoided and work load migration minimized. As a part of the work data collected, pre-processed and linear regression used to calculate threshes hold value. Thresh hold value used to calculate the minimum workload to be maintained by a single machine. The proposed algorithm evaluation performed in open stack shows that present work proves its dynamic nature with more elasticity and less make span time.