No abstract

No abstract

No abstract

This paper addresses a resource selection problem for applications that update data in enterprise grid systems. The problem is insufficiently addressed as most of the existing resource selection approaches in grid environments primarily deal with read-only job. We propose a simple yet efficient algorithm that deals with the complexity of resource selection problem in enterprise grid systems. The problem is formulated as a Multi Criteria Decision Making (MCDM) problem. Our proposed algorithm hides the complexity of resource selection process without neglecting important components that affect job response time. The difficulty on estimating job response time is captured by representing them in terms of different QoS criteria levels at each resource. Our experiments show that the proposed algorithm achieves very good results with good system performance as compared to existing algorithms.

A resource selection problem for asynchronous replicated systems in utility-based computing environment is addressed in this paper. The needs for a special attention on this problem lies on the fact that most of the existing replication scheme in this computing system whether implicitly support synchronous replication and/or only consider read-only job. The problem is undoubtedly complex to be solved as two main issues need to be concerned simultaneously, i.e. 1) the difficulty on predicting the performance of the resources in terms of job response time, and 2) an efficient mechanism must be employed in order to measure the trade-off between the performance and the monetary cost incurred on resources so that minimum cost is preserved while providing low job response time. Therefore, a simple yet efficient algorithm that deals with the complexity of resource selection problem in utility-based computing systems is proposed in this paper. The problem is formulated as a Multi Criteria Decision Making (MCDM) problem. The advantages of the algorithm are two-folds. On one fold, it hides the complexity of resource selection process without neglecting important components that affect job response time. The difficulty on estimating job response time is captured by representing them in terms of different QoS criteria levels at each resource. On the other fold, this representation further relaxed the complexity in measuring the trade-offs between the performance and the monetary cost incurred on resources. The experiments proved that our proposed resource selection scheme achieves an appealing result with good system performance and low monetary cost as compared to existing algorithms.