A Study and Modeling for Improving Grid Service Reliability Using Hive

Abstract

The promise of Grid computing organizations being able to acquire all the power they need for only as long as it is necessary is incredibly compelling. Although grid computing has experienced significant success in bringing productivity gains and cost savings to engineering it has not yet proven its mettle for highly transaction oriented applications. When a program needs to access a remote resource, various types of failures are interleaved in the grid computing background, such as jamming failures, time-out failures, matchmaking failures, program failures, resource failures. This paper presents a grid service reliability model using hive that is used to expand the range of problems that can be solved with grid and host the transaction-oriented applications upon which businesses depend. In the proposed model, client request are handled by ontology, a database to check whether the resources available in the particular server or not. After finding the resource available, hive handles the request and moves the request to remote execution of Condor-G Scheduler. The Condor-G Scheduler executes the request and allocates the resource to the client. Thus the system reliability is afforded.

Keywords: Ontology, Condor G Scheduler, RMS, Hive Computing, Reliability

View PDF