DESIGN OF IMPROVED HYBRID LOAD BALANCING ALGORITHM FOR CLOUD COMPUTING ALGORITHM

Abstract

Cloud computing is one of the leading technological advancements in the 21st century. Cloud computing can be defined as the delivery of a wide range of computing services over the Internet. These services are databases, software, servers, storage, networking, and other resources. Cloud computing provides faster innovation, flexible resources, and economical services. It follows the pay-per-use rule. Load balancing is a method of fairly sharing the workload between all nodes of virtual machines in cloud computing. Some of the most significant issues in cloud computing are task scheduling and load balancing amongst virtual machines as customers access dispersed data centers from various continents. Different algorithms are proposed to tackle this problem, many of which are successful with some limitations. We also proposed a hybrid algorithm based on two existing algorithms. The proposed algorithm takes advantage of the existing features of Throttled load balancing algorithm and ESCE algorithm and improves them to provide better load distribution between virtual machines. In this thesis, an improved hybrid loadbalancing algorithm has been developed. Load distribution is performed to virtual machines based on the properties of the virtual machines. The number of allocated tasks to a virtual machine and its bandwidth are the factors considered when choosing virtual machines. Two virtual machines are picked randomly, and the one with the higher weight is selected for the task. According to the experiments, the proposed algorithm has better performance and efficiency than the three algorithms presented in this study in terms of load distribution and response time relatively based on the property of the virtual machine.