Cost-Deadline Constrained Robust Scheduling of Workflows Using Hybrid Instances in IaaS Cloud

Cloud Computing has gained popularity due to the on-demand resource allocation in the distributed computing environments and provides resources that are dynamically scalable on the “pay as you go” model. Over the past years, Amazon has started providing a new service called EC2 Spot Instances which provides their idle machines on rent in the spot market at a lower cost. Spot instances are the unused virtual machines which are accessible at almost 75% lower price than their on-demand price to perform compute-intensive tasks. Spot instances will end up till the current spot price is less than the user’s bid price. In this paper for the cloud environment, the IaaS Cloud Partial Critical Paths(IC-PCP) algorithm is expanded with the aim of reducing execution costs while still meeting user-defined deadlines. ICPCP schedules the task by finding a computation service which can execute complete critical path before its latest finish time. The work presented in this paper proposes a workflow scheduling algorithm, IaaS Cloud Partial Critical Paths with robustness(IC-PCPR) uses both on-demand and spot instances to reduce the cost of workflow execution while satisfying a user-defined deadline and making system robust that runs on heterogeneous resources. The proposed work is simulated in MATLAB framework and the experimental results based on three scientific workflows show that the ICPCPR performs much better than ICPCP algorithm.