Deploying a Campus Grid: Experience With The Condor Distributed Batch System

Deploying a Campus Grid: Experience With The Condor Distributed Batch System

Paper in proceedings (conference paper)

It is not uncommon now to find research problems that require weeks or months of computation to solve. That’s why we need a computing environment that delivers large amounts of computational power over a long period of time. Such an environment is called a High-Throughput Computing (HTC) environment. On the other hand, the department now owns tens of powerful workstations, which are connected by local area networks and often sit idle for very long periods. For example, workstations at computer labs are used by students just during the tutorials; between and after the tutorials are underutilized. The objective of this work is thus to build a HTC environment which could make use that idle workstations.

It is not uncommon now to find research problems that require weeks or months of computation to solve. That’s why we need a computing environment that delivers large amounts of computational power over a long period of time. Such an environment is called a High-Throughput Computing (HTC) environment. On the other hand, the department now owns tens of powerful workstations, which are connected by local area networks and often sit idle for very long periods. For example, workstations at computer labs are used by students just during the tutorials; between and after the tutorials are underutilized. The objective of this work is thus to build a HTC environment which could make use that idle workstations.

Condor, Grid

Condor, Grid

Začal J, Jan J, Jiřík J, Kolář J, Kubečka L, Peterlík I

19.12.2005

UTIA - Academy of Science of Czech Republic

Praha

International Workshop on Data - Algorithms - Decision Making

1

1