Performance Evaluation of Enhanced Heterogeneous Earliest Finish Time Algorithm for DAG Task Scheduling in Cloud Computing

Abstract

Cloud environment provides enormous measure of computing, storage and networking resources. Managing and providing process resources to extensive number of users and execution of tremendous applications is a challenge in cloud computing. Attaining proficiency and providing fairness to tasks execution, scheduling of tasks is essential.  Task scheduling refers assigning resources to every task and it becomes the essential factor in increasing the performance for the dynamic allocation of resources. The services application storage, network, server and other services can be utilized efficiently and it results reduced makespan and optimum cost. Heuristic techniques are based upon extensive search. Heterogeneous Earliest Finish Time (HEFT) algorithm is a heuristic which arranges tasks by rank function. A greater priority is given to the task with a higher rank and every task is assigned to resources according to its priority. In this research work, an Enhanced HEFT (EHEFT) heuristic algorithm is proposed with a pareto optimization operator called crowding distance to designate priorities to the equal ranked tasks. The proposed method is compared with HEFT, DHEFT, MaxMin, MCT and MinMin and obtains the quantitative measurements for makespan and total cost.