Agent-based Robust Grid Scheduling Framework for High Performance Computing

Abstract

Grid computing is the enabling technology for high performance computing in scientific and large scale computing applications. A number of powerful parallel systems have been built with increasing number of processors and multi-core solutions in order to meet the increased computational demands. Scheduling is a fundamental issue in achieving high performance on computational Grids. A Grid scheduling system must meet the functional requirements of heterogeneous domains (e.g., user domains, application domains, and network domains). Sometimes these requirements conflict with one another. Moreover, a Grid scheduling system must also satisfy non-functional requirements, such as reliability, efficiency (in terms of time consumption), effectiveness in resource utilization, and robustness.Many frameworks have been proposed for Grid scheduling system. Agents are able to behave autonomously, intelligently, learn from environment, and adaptability. Agent-based Grid scheduling is concerned with the design and development of software agents for the resource discovery, job scheduling, resource management, and allocation of computing resources on a computational Grid. This paper presents a new Agent based Grid scheduling framework which possesses high degree of robustness, performance, efficiency, scalability and heterogeneity.