An energy-efficient reliable grid scheduling model using NSGA-II

Abstract

parameters by effectively managing the available grid resources. In general, it is unusual to assume a single criterion for the decision making in a grid system which is governed by various related factors. The scheduling problem in the grid system may have several conflicting objectives, e.g., makespan, reliability, energy consumption, cost, security, etc., which need to be optimized simultaneously [1][2][3][4].The outcome values of different objectives, viz. execution time of any job, the energy consumption by the job and the reliability offered by the system to the job may vary from one allocation to another [5]. The distributed nature of grid resources and the dynamic nature of the grid workload make the makespan of the job unpredictable. For makespan minimization, it is warranted to effectively deal with the load balancing as well [6]. Further, the heterogeneous hardware components of the grid are more prone to failure and therefore reliability of such systems decreases with the increase in the number of geographically distributed resources [7]. The heterogeneity of the system demands varying energy consumption based on their hardware profile [8]. It necessitates choosing those nodes/machines that are able to execute the job quickly, reliably and with least energy consumption.Since these objectives are conflicting in nature, any approach adopted to optimize one objective may not result in the overall optimized solution with respect to other objectives. All the conflicting objectives need to be addressed simultaneously and the best way for this is to have the trade-off solutions with all the objectives [9, 10]. A good solution corresponding to all the objectives is desired. This solution may not offer the best of all the objectives, but certainly better when seen as a trade-off among the objectives. This approach is referred as multiobjective optimization technique [5]. As such, job scheduling in the grid is an NP-hard problem and traditional Abstract Job scheduling in computational grid is a complex problem and various heuristics and meta-heuristics have been proposed for the same. These approaches usually optimize specific characteristic parameters while allocating the jobs on the grid resources. Many a times, it is desired to optimize multiple parameters during job scheduling. Nondominated sorting genetic algorithm (NSGA-II) has been observed to be the best meta-heuristic to solve such multiobjective optimization problem. The proposed work applies NSGA-II for job scheduling in computational grid with three conflicting objectives: maximizing reliability of the system for job allocation, minimizing energy consumption and balancing the load on the system. Performance study of the proposed model is done by simulating it on some real data. The result indicates that the proposed model performs well with multiple objectives.

A Comprehensive Review on NSGA-II for Multi-Objective Combinatorial Optimization Problems

“…The proposed NSGA-II outperformed the two state-of-the-art approaches, the GA and the greedy algorithm, in finding the best compromised optimal schedules. Kaushik & Vidyarthi [180] proposed a multi-objective resource allocation model for computational grid scheduling using a dynamic resource allocation scheme. Further, NSGA-II used to optimize this model proved better than GA as it offered alternative trade-off solutions instead of a single best compromise solution.…”

Analysis of Effective Load Balancing Techniques in Distributed Environment

“…Kaushik and Vidyarthi [36] consider various parameters for effective job scheduling and resource allocation. The presented model selects the best cluster in terms of increased system reliability and reduced energy consumption and balances the system load efficiently.…”

Performance aware scheduling considering resource availability in grid computing

“…The total makespan of a grid computing environment is defined as the largest makespans of grid resources. The makespan of a resource is the time slot between the start and completion of a sequence of jobs assigned to that resource [6,[10][11][12][13][14][15][16][17][18][19]. Actually, grid is considered as a high throughput system, and minimizing the total makespan of the grid increases the throughput of the environment accordingly.…”

“…Since decision about the assignment of jobs to resources and finding the best match between jobs and resources is an NP-complete problem [6,[10][11][12][13][14][15][16][17][19][20][21][22], we use Simulated Annealing (SA) meta-heuristic [11,16,[19][20][21][22][23] to propose a job scheduling algorithm to solve the problem. SA is a generalization of the Monte Carlo method for statistically finding the global optimum for multivariate functions.…”