Abstracto

Improving the Performance of Condor for Dag- Based Tasks through Cache Memory in Grid Environments

N. Purandhar, N.Usharani

Grid application consists of one or more tasks. These tasks are generally divided into two types such as Independent task and Workflow tasks. A number of resource brokers that act as middleware such as Condor, Nimrod-G etc are used in grid based application. Since, workflow tasks are not easily executed with these architecture, we slightly modified the DAG-Condor architecture to suit workflow tasks. This paper focuses on cache management system in DAGCondor resource broker. By adding cache to store intermediate results and these results also stored in central database which contains old results. The overhead of read and writes to central database is reduced. It improves the performance, throughput of the system.

Descargo de responsabilidad: este resumen se tradujo utilizando herramientas de inteligencia artificial y aún no ha sido revisado ni verificado.