Resource-aware meta-computing

Meta-computing, is an increasingly popular and useful method of obtaining resources to solve large computational problems. However, meta-computer environments pose a number of unique challenges, many of which have yet to be addressed effectively. Among these are dynamicism in both applications and environments, and heterogeneity at several different levels. This chapter discusses current approaches to these problems, and uses them in the Active Harmony system as a running example. Harmony supports an interface that allows applications to export tuning alternatives to the higher-level system. By exposing different parameters that can be changed at runtime, applications can be automatically adapted to changes in their execution environment caused by other programs, the addition or deletion of nodes, or changes in the availability of resources like communication links. Applications expose not only options, but also expected resource utilization with each option and the effect that the option will have on the application's performance. We discuss how this flexibility can be used to tune the overall performance of a collection of applications in a system.