Title:Energy Efficient Reliability Aware Workflow Scheduling in Cloud Computing
Volume: 7
Issue: 3
Author(s): Nidhi Rehani and Ritu Garg*
Affiliation:
- Computer Engineering Department, National Institute of Technology, Kurukshetra, Haryana,India
Keywords:
Cloud computing, reliability, multi objective workflow scheduling, energy-efficiency, Monte
Carlo Simulation, green computing.
Abstract: Background & Objective: In today’s scenario, workflow scheduling algorithms require multiple
conflicting goals to be optimized. Optimal makespan (schedule length) for the workflow application
is the most important criterion to be optimized to achieve desirable performance. Reducing energy
consumption for high performance computing requirements is extremely important to control the rapidly
growing demand for computation power. This results in a decrease of the operational cost and carbon-
dioxide emissions to the environment. Moreover, the computer processors in a heterogeneous environment
are not failure free. Any kind of failure can be critical for an application. In this paper, we propose
a multi-objective workflow scheduling algorithm in cloud computing – ERAWS, which optimizes
three conflicting criterions: makespan, reliability of task execution and energy consumption. We validate
and analyze the performance of our algorithm by using the CloudSim toolkit to simulate the cloud
environment. We compare the performance of our algorithm with HEFT and ECS, using randomly generated
task graphs and task graphs for real world problems like Gaussian Elimination and Fast Fourier
Transformation to represent workflow applications.
Conclusion: The simulation results show that the proposed ERAWS algorithm is significantly better
than the considered algorithms in terms of makespan, reliability and energy consumption in real world
scenarios where reliability and energy consumption are important issues.