An integrated classAd-latent semantic indexing matchmaking algorithm for globus toolkit based computing grids

Abstract

The effective and efficient computing resource allocation is a critical issue in the challenging achievement of realistic, high space and temporal resolution and computing time affordable weather, marine and soil flooding simulation and forecast. Resource allocation relies on matchmaking tool, which must be flexible and straightforward to be deployed, configured and used. Using the Globus Toolkit 4 and our resource broker service, we compared the ClassAd matchmaking algorithm with our matchmaking algorithm and then we integrated the two algorithms in order to avoid their drawbacks. © 2008 Springer-Verlag Berlin Heidelberg.