期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2015
卷号：3
期号：2
DOI：10.15680/ijircce.2015.0302177
出版社：S&S Publications
摘要：Most of the existing grid scheduling algorithms allocates jobs to the resources and when they fails itsimply reschedules the job to another optimal resource. So in this method, the failed resource is not attempted torecover instead it is left as it was. This leads to the affects the maximum utilization of resources, also it causes excessbandwidth to transfer job from one resource to the other. As the failed resource is not recovered or it is not taken care,in due course the failure rate of the resource is increased and job becomes unfit for job allocation.In the proposed system, resource allocation is done using a conventional algorithm, but when the job fails due toprocess (which is currently executing the job) failure it is attempted to recover. By doing so the resource failure due toprocess failure is recovered. While the resource fails it is detected for process failure. When it detected for processfailure it attempted to recover. When we detect the process failure, the process is attempted to recover with anotherprocess in the same resource. When the process is unrecoverable we complete the job with another process in the sameresource. Thus in our method, when process failure happens the rescheduling is avoided.
关键词：Ontology; personalization; knowledge base; user profiles; web information gathering
