Execution Time Based Load Distribution over Virtual Machines in Cloud Environment

摘要

Cloud computing allows the end users to access a set of distributed hosted resources like storage, CPU, applications (both software and hardware) and services over the internet in an on-demand basis. The interlinked servers of cloud network are responsible to deliver these services to the users. Load balancing is an inevitable challenging procedure associated with the computing environment of cloud. The dynamic load (or tasks) has to be distributed across multiple virtual machines or computers in an efficient manner so as to achieve maximum resource utilization and improved response time of submitted tasks. Existing load balancing algorithms attempt to transfer tasks from overloaded machine to underloaded machine thereby avoiding a situation of creating idle or low loaded virtual machines. This paper presents the concepts and techniques of load balancing in cloud environment and proposes a new load balancing procedure based on two QoS factors such as priority and execution time and it maintains all virtual machines remain balanced all times and thus achieving maximum resource utilization and good response time. The objective of this paper is to analyze the factors affecting an efficient load balancing procedure in cloud environment.