The problem of efficient assignment of resources to perform a given bag-of-tasks in a distributed computing environment has been extensively studied by research communities. To develop an efficient resource assignment mechanism, this paper focuses on a particular type of resource intensive tasks and presents a bi-level decision-making architecture in a grid computing environment. In the proposed architecture, the higher decision-making module has the responsibility to select a partition of resources for each of the tasks. The lower decision-making module uses Integer Linear Programming based algorithm to actually assign resources from this selected partition to a particular task from the given set of tasks. This paper analyzes the performance of the proposed architecture at various workload conditions. This architecture can be extended for other types of tasks using the concepts presented.

Additional Metadata
Keywords Distributed systems, Dynamic adaptation, File transfers, Grid computing, High performance computing
Citation
Ahmad, I. (Imran), & Majumdar, S. (2007). Efficient management of grid resources using a Bi-level decision-making architecture for "processable" bulk data.