Impatient Task Mapping in Elastic Cloud using Genetic Algorithm

Abstract

Problem statement: Task scheduling is the main factor that determines the performance of any distributed system. Cloud computing comes with a paradigm of distributed datacenters. Each datacenter consists of physical machines that host virtual machines to execute customers’ tasks. Resources allocation on the cloud is different from other paradigms and the mapping algorithms need to be adapted to the new characteristics. This study takes the problem of immediate task scheduling under an intercloud infrastructure using a genetic algorithm. An impatient task needs to be scheduled as soon as it enters the system taking into account the input and output files location and its QoS requirements. Approach: This study proposes an algorithm that can find a fast mapping using genetic algorithms with “exist if satisfy” condition to speed up the mapping process and ensures the respecting of all task deadlines. Cloudsim simulator was used to test the proposed algorithm with real datasets collected as a cloud benchmark. Mapping time and makespan are the performance metrics that are used to evaluate the proposed system. Results: The results show an improvement in the proposed system compared to MCT algorithm as illustrated throughout the study. Conclusion: Batch mapping via genetic algorithms with throughput as a fitness function can be used to map jobs to cloud resources.