Solving resource provisioning in cloud using GAs and PSO

Cloud Computing can be used as a buzzword for the big turn to the world where computing resources are provided over the internet. This paper presents two phase approach to solve the cloud resource provisioning problem from consumer's perspective. To minimize the budget, consumer must find exact number of resources required and select proper resource purchasing plan. First phase deals with minimization of number of instances of virtual machine required to execute workflow tasks which belongs to category of NP problem. The second phase is resource subscription phase which determine resource purchasing plan based on tipping point calculation. The primary objective of this paper is to compare performance of Genetic Algorithms (GAs) and Particle Swarm Optimization (PSO) for tasks to virtual machine mapping. GAs and PSO reported good solutions with more than 60% VMs utilization. GAs reported better solutions than PSO for more than 60% times with less number of iterations.