Compromise Scheme for Assigning Tasks on a Homogeneous Distributed System

—We consider the problem of assigning tasks to homogeneous nodes in the distributed system, so as to minimize the amount of communication, while balancing the processors' loads. This issue can be posed as the graph partitioning problem. Given an undirected graph G=(nodes, edges), where nodes represent task modules and edges represent communication, the goal is to divide n, the number of processors, as to balance the processors' loads, while minimizing the capacity of edges cut. Since these two optimization criteria conflict each other, one has to make a compromise between them according to the given task type. We propose a new cost function to evaluate static task assignments and a heuristic algorithm to solve the transformed problem, explicitly describing the tradeoff between the two goals. Simulation results show that our approach outperforms an existing representative approach for a variety of task and processing systems. I.INTRODUCTION THE cluster systems are not only provide facilities for utilizing resources and/or data at remote sites, but also enhance system performance and reliability with the multiplicity of processors and communication paths. Distributed applications range from large data base installations where the computing load is distributed for organizational efficiency, to small signal-processing systems where computation must be done very fast in a real-time environment. There are, however, many problems to be resolved before realizing the potential of a distributed system, such as file and task assignment [6]. To alleviate part of these problems, we consider the problem of assigning tasks in cluster computing systems. Specifically, we will deal with static assignment of a given task to the processors in a distributed system which neither requires nor precludes the subsequent dynamic migration of the task. In particular, we are interested in developing centralized task assignment algorithms using global knowledge of the task and system characteristics. These task assignment algorithms attempt to assign each task to the processors so as to achieve such goals as the minimization of inter-processor communication, good load balancing among the processors, a small response time for the task, and efficient utilization of system resources in general. This statement of the static task assignment problem is also useful for assignment of tasks to fewer processors than the distributed algorithm was initially designed for, at the time of initial assignment or dynamically at execution time due to node failure, node withdrawal, or phase transitions in the distributed computation [7]. In this paper, we propose a …