Multiobjective Workload Scheduling in Distributed Datacenters

Cloud computing is now a popular field for indulging infrastructure–as-a-service (IaaS). IaaS defers, on demand virtualized computing resources through datacenters. Cloud users acquire computing infrastructure for the execution of application workloads. However due to increasing data volumes and parallel processing of workloads on limited resources of single data center, workloads are shifted to the other datacenters over the peak time. Therefore scheduling of workloads in distributed datacenter is a challenge as each workload executes across disbursedenvironment having heterogeneous resources and also requires coordination among the datacenters. In this paper, we propose MO-GRASP (Multiobjective Greedy Randomized Adaptive Search Procedure) based heuristic to solve the multi-objective workload scheduling problem. The model objectives are minimization in workload completion time i.e. datacentermakespan and energy consumptionwhilemaximizing the average datacenter utilization. The simulation results show that the proposed algorithm outperforms as compared to other Multiobjective algorithm.