Publication Date:
2010
abstract:
We present a distributed matchmaking methodology based on a two-level (low-level and application-level) benchmarking, that allows
the specification of both syntactic and performance requirements.
In particular, we point out how the use of application-level benchmarks gives a more accurate characterization of resources, so enabling a better exploitation of Grid power.
The proposed methodology relies on the use of standard description languages at both application and resource sides, to foster
interoperability.
Moreover, the proposed tool is independent of the underlying middleware, and its distributed structure supports scalability.
Iris type:
01.01 Articolo in rivista
Keywords:
Grid platforms; Benchmark-driven resource selection; Interoperability; Job submission language extensions
List of contributors:
Galizia, Antonella; Quarati, Alfonso; D'Agostino, Daniele; Clematis, Andrea; Corana, Angelo
Published in: