A Hybrid Approach for Time Efficient and Reliable Workflow scheduling in Cloud Computing

Abstract

Nowadays, scientific and business applications are becoming quite complex all around the world. These large scale problems require computing power that goes beyond the capabilities of a single machine. The data and computation requirements of these problems demand a high performance computing environment such as a cluster, a grid or a cloud platform in order to be solved in a reasonable amount of time. Workflow provides an easy and efficient way to implement these applications. Many companies and researchers are opting to implement the workflow in cloud computing. In order to efficiently execute workflows and utilize the cloud resources in an appropriate way, scheduling policies should be implemented. Most of the workflow applications hosted by cloud are real time applications which require high reliability and fast execution. A number of fault can occurs in cloud due to hardware faults, software faults, network faults and so on. So, there is need of methods for providing the reliable execution of the applications which are efficient in terms of time. In this thesis, we have proposed a hybrid approach for time efficient and reliable workflow scheduling in cloud as a combination of two existing state-of-the-art heuristics which considers the parameters of reliability and time while scheduling the tasks. Using simulation, we have compared the performance of our proposed MCTR algorithm with the existing reliable scheduling algorithm. Experiment results show that proposed algorithm gives better results which optimize makespan of workflow while providing reliable execution