International Journal of Computer Applications (0975 – 8887) Volume 168 – No.9, June 2017 49 Scheduling Tasks in Heterogeneous System using Load Balancing Algorithm Pratiksha Patil Dr. D.Y.Patil School of Engineering and Technology, Lohagoan, Pune Roshani Ade, PhD Dr. D.Y.Patil School of Engineering and Technology, Lohagoan, Pune ABSTRACT Innovative idea of distributing the tasks to their best processor to reduce the execution time of task by using various scheduling techniques is given in this paper. This paper presents hybrid scheduling techniques which provide better solution of scheduling task that means combination of different scheduling provides better performance without degrading the result quality. Scheduling algorithms such as MinMin+, MaxMin+ and Sufferage+ are suitable for overcomes the drawback of previously used scheduling methods such as MinMin, MaxMin and Sufferage as well as scheduling in this paper provides better complexity as compare to previous scheduling methods. This scheduling are also suitable for heterogeneous environment more effectively to execute different set of task on different processors with different configurations. To get better the show of the existing system we have to improve existing algorithm with the load balancing.So same load should overloaded to all processors.The future algorithm have implement with detailed pseudocodes. Keywords Task Scheduling, MinMin, MaxMin, Sufferage, Standard Deviation, Load Balancing. 1. INTRODUCTION Distribution of large application into task for faster processing is one of the important process in the area of distributed systems. Although many types of resources can be shared and used in a distributed system, usually they are accessed through an application running in the network. Normally, an application is used to define the piece of work of higher level in the system. An application can generate several tasks, which in turn can be composed of subtasks; this system is responsible for sending each subtask to a resource to be solved. It performs an important step of mappings task to different machines based on the expected execution time. Normally an application is used to define the piece of work of higher level in heterogeneous environment. Since this application can generate several number of jobs that can be divided into subtasks and provided to different processors that should get completed within minimum time so that the processor use can be made to assign different task. Makespan is one of the most important term in case of mapping task to their processors using different scheduling. Makespan is nothing but turnaround time that is maximum of completion time. An optimal schedule will be the one that minimizes the makespan [1, 2]. Large numbers of task scheduling algorithms are available to minimize the makespan. All these algorithms try to find resources to be allocated to the tasks which will minimize the overall completion time of the jobs. The simple well-known existing algorithms used for scheduling are MinMin and Maxmin and sufferage. These algorithms work by considering the execution and completion time of each task on the each available grid resource. Scheduling is considered to be an important issue in the current distributed system existing algorithms used for scheduling are Min-Min and Maxmin and sufferage. These algorithms work by considering the execution and completion time of each task on the each available grid resource. Scheduling is considered to be an important issue in the current distributed system scenario. The demand for effective scheduling increases to achieve high performance computing. Typically, it is difficult to find an optimal resource allocation which minimizes the schedule length of jobs and effectively utilize the resources. The three main phases of scheduling are resource discovery, gathering resource information and job execution. The choice of the best pair of jobs and resources in the second phase has been proved to be NPcomplete problem. The existing scheduling algorithms provide the various techniques for assigning different task to different resources with minimum completion time. These existing scheduling algorithms can be divided into two classes that are Online mode and Batch mode scheduling. In online mode, a task is assigned to processor on its arrival to scheduler [1]. Wherein Batch mode scheduling tasks are not assigned to processor immediately instead they are collected in to set of tasks also called as Metatask that are examined for assigning at prescheduled times to different processors also called as mapping events. Since in this system, batch mode is used in very efficient way for mapping different independent tasks to processors. Also these existing algorithms can be applied for heterogeneous environment effectively. The proposed system in this work contains various scheduling methods along with hybrid technology such as Minmin+, Maxmin+, and Sufferage+. Hybrid technology involves combination of different scheduling methods to overcome disadvantages of minmin and maxmin. Overall, the scheduling algorithms aim to minimize the idle time and makespan of tasks. This paper also involves the concept of Load Balancing [2, 6], wherein, once scheduling of task is done using some scheduling the load balancing algorithm will take place to reschedule the task to utilize all the resources in the heterogeneousenvironment [5]. Each of this scheduling provides better performance and also decreases time complexity without degrading the solution quality. To avoid the drawbacks of the existing scheduling algorithm, the proposed system algorithms are being used to enhance the system performance. Every one of the issues talked about in those techniques are taken and dissected to give a more powerful schedule. The calculation proposed in this paper beats every one of those calculations both in wording of makespan and load balancing. In this way a superior load