An Agent-based Service Discovery Algorithm Using Agent Directors for Grid Computing

Abstract- Grid computing has emerged as a viable method to solve computational and data-intensive problems which are executable in various domains from business computing to scientific research. However, grid environments are largely heterogeneous, distributed and dynamic, all of which increase the complexities involved in developing grid applications. Several software has been developed to provide programming environments hiding these complexities and also simplifying grid application development. Since agent technologies are more than a ten-year study and also because of the flexibility and the complexity of the grid software infrastructure, multi-agent systems are one of the ways to overcome the challenges in the grid development. In this paper, we have considered the needs for programs running in grid, so a three-layer agent-based parallel programming model for grid computing is presented. This model is based on interactions among the agents and we have also implemented a service discovery algorithm for the application layer. To support agent-based programs we have extended GridSim toolkit and implemented our model in it