Publication

Abstract : Resource management during the cloud computing process is one of the prominent issues currently under research. The reason behind this is to procure an optimal utilization of these resources with the least cost and time. Allocation and scheduling of users’ tasks over the servers is an NP-hard problem, which can be solved by applying various heuristic and meta-heuristic approaches. With this motive, the paper has analyzed the performance of two meta-heuristic approaches, i.e., Genetic Algorithm (GA) and Flower Pollination Algorithm (FPA) in scheduling the tasks over virtual machines (VM). Tasks to VMs allotment is mainly responsible for the utilization of virtual resources which are actually running over physical machines (PM) or servers. The prime parameter considered in the research is makespan which indirectly measures the consumption of resources for performing a set of tasks. Simulation for the research is performed for a defined set of tasks that need to be allocated over a particular number of servers using the cloudsim tool. Results are measured and analyzed for both the algorithms applied in the scheduling process and also a default approach of first come first serve (FCFS) is compared with the overall results. Analyses prove the GA approach overperforms the FPA in reducing the makespan in various configurations considered.