Multi-Objective Fault Tolerance Model for Scientific Workflow Scheduling on Cloud Computing

Abstract

– Cloud computing is used for large-scale applications. Therefore, a lot of organizations and industries are moving their data to the cloud. Nevertheless, cloud computing might have maximum failure rates because of the great number of servers and parts with a high workload. Reducing the false in scheduling is a challenging task. Hence, in this study, an efficient multi-objective fault detector strategy using an improved Squirrel Optimization Algorithm (ISOA) in cloud computing is proposed. This method can effectively reduce energy consumption, makespan, and total cost, while also tolerating errors when planning scientific workflows. To increase the detection accuracy of failures, the Active Fault Tolerance Mechanism (PFTM) is used. Similarly, the reactive fault tolerance mechanism (RFTM) is used for processor failures. The efficiency of the proposed approach is analysed based on various measurements and performance compared to other approaches.