FUZZY LOGIC CONTROLLED RESOURCE ALLOCATION FOR EFFICIENT LOAD BALANCING IN CLOUD COMPUTING ENVIRONMENT

This paper address the problem of managing cloud system, consisting a set of virtual machines (VMs), operating under dynamic workloads conditions. The objective of the paper is to find the best workload-VM pair in such a way to give a guarantee on his Quality-of-Services (QoS) and, at the same time, to minimize the energy consumption of the physical infrastructure running them. For that we propose a fuzzy controller, which is able to allocate the best VMs to each cloud application in such a way that the minimum amount of physical capacity needed to meet its QoS requirements. In this way the VMs are dynamically selected hence the number of physical resources that needed to be active at any given instant of time is reduced in comparison to the statically provisioned systems. This reduces the energy requirement to run a given cloud workload. We implemented a prototype of our controller on CloudSim, and tested it over different load conditions in which we compare the proposed technique against, state-of-the-art techniques. Experimental results show that proposed technique outperforms state-of-the-art techniques in terms of QoS and the amount of physical resources required.