有服务质量保证的数据密集型网格应用管理研究

在网格中数据密集型应用的管理由于需要综合考虑多种资源属性而变得非常复杂,本文针对该问题提出了一种有服务质量保证的数据密集型网格应用管理方法,该方法通过两个方面来保证服务质量,首先是通过基于数据可靠性的副本管理策略提高数据的可用性,其次采用基于任务截至期限的任务调度策略来动态调度任务.最后,给出模拟试验测试以及实际网格平台与实际应用的测试,测试结果表明本文提出的方法较其它方法有很大提高.

Data-Intensive Applications Management Framework with QoS Guarantee in Grid Environments

Effectively management of data-intensive applications in grid environments is challenging, due to a need to consider for both computational resources and data storage resources. This paper proposes a management framework for such applications with QoS Guarantee. This framework includes two parts, one is replica management based on availability of data, and the other is deadline-based job scheduling model. In the end, a simulation evaluation and two real applications evaluation over HUSTGrid are presented, the results of the evaluations show that this framework is suitable for data-intensive applications, and the performance of this scheduling method is better than the related methods.