Prediction based Load Balancing and VM Migration in Big Data Cloud Environment

Abstract

In Big Data Cloud atmosphere, the cloud service provider (CSP) offers amenities to the customer with the accessible virtual cloud sources. Investigators have been provided more consideration towards the harmonizing of the load, as it has a complete impact on the system act. In this paper, Prediction based Load Balancing and Virtual Machine (VM) Migration (PLBVM) algorithm is designed for Big data cloud environments. In this algorithm, the future loads of each server are estimated. If the estimated future load is greater than an upper bound or less than a lower bound, then it indicates unbalanced load, so that VM migration is triggered. In VM migration, the VMs with minimum migration time and sufficient resources are selected. Then the task execution continues in the migrated VMs. By experimental results, it is shown that PLBVM achieves lesser response delay and execution time, among the other approaches.