A budget constrained scheduling algorithm for executing workflow application in infrastructure as a service clouds

Abstract

Cloud computing technology, which is a new model of service provisioning in distributed systems, has been raised as a way to execute workflow applications. To profit from this technology for executing workflow applications, it is necessary to develop workflow scheduling algorithms that consider different QoS parameters such as execution time and cost. Therefore, in this paper, we focus on two criteria: makespan (completion time) and execution cost of workflow application and propose a scheduling algorithm named CB-DT (Constrained Budget-Decreased Time) which aims to create a schedule that decreases the makespan while satisfying the budget constraint of the workflow application. In the proposed algorithm, the ideas of back-tracking heuristic and scheduling of critical and non-critical tasks are combined together. In order to have smaller makespan, the proposed algorithm tries to select faster and more expensive machines for critical tasks as much as possible using the back-tracking method. Moreover, it tries to schedule non-critical tasks on the low cost machines as far as possible without increasing the makespan. The proposed algorithm is evaluated by a simulation process using WorkflowSim which is based on CloudSim. To evaluate the proposed algorithm, the results of proposed algorithm are compared with the results of IC-Loss (IaaS Cloud-Loss), BHEFT (Budget constrained Heterogeneous Earliest Finish Time) and BDHEFT(Budget and Deadline Constraint Heterogeneous Earliest Finish Time) algorithms. The results showed that the proposed algorithm performs better than IC-Loss, BHEFT and BDHEFT algorithms in most cases.