Harnessing the Capacity of Computational Grids for High Energy Physics

Abstract

By harnessing available computing resources on the network, computational grids can deliver large amounts of computing capacity to the high energy physics (HEP) community. Supporting HEP applications, which typically make heavy memory and I/O demands, requires careful co-allocation of network, storage, and computing resources. The grid manager must ensure that applications have the necessary resources to run efficiently while respecting the usage policies imposed by the owners of those resources. The Condor research group is engaged in an effort to develop effective co-allocation mechanisms for computational grids. We present two of these mechanisms, checkpoint domains and file system domains, which attack the wide-area network bottleneck by providing local access to checkpoint and data storage. Keywords: grid computing, network scheduling, high throughput computing 1 Introduction Computational grids [1] have the potential to deliver large amounts of computing capacity to t...