Fast networks have made it possible to aggregate distributed CPU, memory, and storage resources into Computational Grids. The goal is to provide the potential for application performance superior to that achievable on any single system. However, achieving such performance on the Grid for parallel applications whose performance is highly dependent upon the efficient coordination of their constituent components.

Currently, to achieve a performance-efficient implementation on a distributed heterogeneous system, the performance-seeking end-user or application developer must select a potentially efficient configuration of resources based on load and availability, evaluate the potential performance on such configurations based on their own performance criteria, and interact with the relevant resource management systems in order to implement the application. Our goal is to provide mechanisms and paradigms that perform scheduling and decision making on behalf of the user.

Part of the AppLeS project focuses on developing Application-Level Scheduling agents to provide a mechanism for scheduling individual applications at machine speeds on production heterogeneous systems. AppLeS agents utilize the Network Weather Service (NWS) to monitor the varying performance of resources potentially usable by their applications. Each AppLeS uses static and dynamic application and system information to select viable resource configurations and evaluate their potential performance. AppLeS then interacts with the relevant resource management system to implement application tasks. Once it contains an embedded AppLeS agent, the application becomes self-scheduling

Another effort that is part of the AppLeS project is the development of AppLeS templates. Templates are stand-alone software projects that perform automatically scheduling and deployment tasks for classes of structurally similar applications. Templates builds on the expertise gained while developing AppLeS agents, and the main focus is re-usability for several applications. Two template efforts are underway withing the GRAIL: the Parameter Sweep Template and the Master Slave Template.