Energy Efficient Load Balancing in Computational Grid

Abstract

Computational grid is a pool of computational resources which handles large computations and consumes enormous energy. Load balancing and energy consumption are two main performance parameters of the computational grid. Since performance of a computational grid is highly susceptible to variations in load and energy, therefore it is imperative to pay attention to these aspects in grid environment. In this thesis we proposed an energy aware dynamic load balancing strategy for tasks scheduling in grid referred to as Energy Efficient Load Balancing (EELB) in computational grid. EELB combined energy efficiency and load balancing paradigm to improve the utilization and to reduce the energy consumption of the resources. EELB considered a scenario where users require services which are either time or budget constraint and based on user’s requirements the tasks are allocated to the grid resources for execution. The simulations were carried out using GridSim environment and the results were compared with the Improved Load Balancing on Enhanced GridSim with Deadline Control (IEGDC) method. The performance indices such as: finished gridlets, unfinished gridlets, resubmitted times and response time, along with proposed parameters such as: cost, energy consumption and energy saved were used for the comparison of both methods. The simulation results showed that EELB outperformed the published IEGDC method by 28% in terms of finished gridlets, 35% by cost and 38% in energy consumption.