International Journal of Computer Applications (0975 – 8887) Volume 86 – No 14, January 2014 1 Improve Performance of Load Balancing using Artificial Bee Colony in Grid Computing Deepika Nee Miku M-Tech student of Department of Computer Science and Application Maharshi Dayanand University,,Rohtak, Haryana-124001 Preeti Gulia Department of Computer Science and Applications Maharshi Dayanand University, Rohtak, Haryana- 124001 ABSTRACT Grid computing is a novel approach which solves the load balancing problems in scientific, engineering and research area. Load Balancing is a technique that can be used to improve resource utilization, to reduce MAKESPAN and to minimize number of failures. In grid environment, different algorithm for resources and data distribution is used to increase the performance and efficiency of load balancing. In grid environment Static threshold and PSO are used for load balancing. Static (fixed) threshold i.e. 3 is used for data transfer from source node to server node. Then, using PSO for data transferring that is better than, static threshold. Artificial Bee Colony Algorithm (ABC) is an optimization algorithm based on the intelligent foraging behavior of honey bee swarm. In this paper, propose a new load balancing algorithm using Artificial Bee Colony(ABC) for obtaining minimum makespan and less number of failure then, obtained results are presented and compared with static threshold and PSO. Keywords Grid Computing, Load Balancing, PSO, ABC 1. INTRODUCTION Grid computing is the combination of computer resources from multiple administrative domains applied to achieve a goal, it is used to solve scientific, technical or business problem that requires a great number of processing cycles and needs large amounts of data. Grid computing is a distributed infrastructure which allows large scale resource sharing and system integration. Load balancing is the most important factor to improve the efficiency and performance of multiple nodes in grid based distributed environment. Workload distribution is carried out in such a way that a set of independent tasks are distributed among all the computing nodes of the grid so that the jobs are uniformly distributed and none of the nodes are overloaded or under loaded [1] . In this paper, a new load balancing algorithm is developed using ABC and compare the simulation results of static threshold, PSO and ABC. The rest of the paper is organized as follows. Section 2, describes related work of various swarm intelligence algorithms for load balancing. Section 3, presents importance of load balance in grid environment. Section 4, describes the concept of swarm intelligence techniques PSO and ABC algorithms. Section 5, proposes a new load balancing algorithm using ABC. Section 6, shows the generated Experimental results in the form of table and graphical presentation of static threshold, PSO and ABC based on MAKESPAN and number of failures. Experimental results are simulated in MATLAB simulation tool. Finally, conclusion is discussed in Section 7, with suggestions for future work. 2. RELATED WORK Since the last two decades, swarm intelligence (SI) has been used into many research areas because of its unique behavior which is inherent from the social insects. Swarm intelligence algorithms have been used by scientists, researchers, engineers for loads balancing of data and various optimization problems. Sowmya Suryadevevra et al., [1] proposed an Ant Colony Optimization algorithm for load balancing in grid computing is proposed which will decide the best resource to assign the jobs based on resources capacity At the same time balance the load of entire resources on grid and. The aim of this ACO is to get high throughput and thus increase the performance in grid environment. T. Kokilavani et al., [2] presents a Load balanced task scheduling which is a very important problem in complex grid environment. Author proposed Load Balanced Min-Min (LBMM) algorithm that reduces the makespan and increases the resource exploitation. The proposed method has two-phases. In the first phase, the traditional Min-Min algorithm is executed and in the second phase the tasks are rescheduled to use the unutilized resources effectively. Manish Gupta et al., [3] describes Artificial Bee Colony Algorithm (ABC) optimization algorithm based on the intelligent foraging behavior of honey bee swarm. For assigning the jobs in the system in a manner that optimized the overall performance of the application. Authors proposed, an Efficient artificial bee colony (ABC) algorithm, where they have used additional mutation and crossover operator of Genetic algorithm (GA) in the classical ABC algorithm for solving the job scheduling problem with the criterion to decrease the maximum completion time. C.Kalpana et al., [4] proposed an algorithm namely max min Particle Swarm Optimization with load balancing techniques with the comparison of Swarm Intelligence Algorithms like Ant Colony Optimization. This algorithm is based on the task scheduling in grid environment. They are calculating the QoS constraints to the PSO and ACO like, makespan, Cost and Deadline. Finally, they had balanced the load to the particle swarm optimization, and compared with all other algorithms. Lei Zhang et al., [5] adopted a heuristic approach based on Particle Swarm Optimization (PSO) algorithm to solving task scheduling problem in grid computing. Each particle is represented a possible solution and the position vector is transformed from the continuous variable to the discrete variable. The main aims to generate an optimal schedule so as to get the minimum completion time while completing the tasks. Mr. P.Mathiyalagan et al., [6] proposed an algorithm, which based on PSO for task scheduling problem on computational grids. Task scheduling algorithms based on PSO algorithm applied in computational grid environment. The aims were generating an optimal schedule so as to complete the tasks in a minimum time as well as utilizing the resources in an efficient way. T.Kokilavani et al., [7] proposed a new approach based on the argentine ant’s behavior. The traditional methods try to reduce the overall response time by giving an optimized schedule. But they fail