WORKFLOW USING PARALLEL CAT SWARM OPTIMIZATION: CASE STUDY

Abstract

Cloud Computing has emerged as a service model that enables on-demand network access to a big number of to be had virtualized assets and packages with a minimal management attempt and a minor price. The spread of Cloud Computing technologies allowed coping with complicated applications inclusive of Scientific Workflows, which consists of a set of in depth computational and information manipulation operations. Cloud Computing facilitates such Workflows to dynamically provision compute and garage sources vital for the execution of its duties thanks to the pliancy asset of these sources. However, the dynamic nature of the Cloud incurs new challenges, as a few allocated sources may be overloaded or out of get right of entry to all through the execution of the Workflow. Moreover, for records extensive duties, the allocation method ought to take into account the records placement constraints for the reason that facts transmission time can increase considerably in this situation which implicates the growth of the general crowning glory time and fee of the Workflow. Likewise, for intensive computational responsibilities, the allocation strategy should recollect the sort of the allocated digital machines, more specially its CPU, reminiscence and network capacities. Yet, a vital assignment is the way to efficaciously agenda the Workflow tasks on Cloud sources to optimize its ordinary nice of provider. In this paper, we endorse a QoS conscious algorithm for Scientific Workflows scheduling that goals to enhance the overall pleasant of carrier (QoS) through considering the metrics of execution time, statistics transmission time, value, assets availability and facts placement constraints. We extended the Parallel Cat Swarm Optimization (PCSO) algorithm to put in force our proposed method. We tested our set of rules within two sample Workflows of different scales and we compared the effects to those given by the usual PSO, the CSO and the PCSO algorithms. The outcomes show that our proposed algorithm improves the overall great of provider of the tested Workflows.