Issue No.05 - May (2008 vol.19)
Multiple applications that execute concurrently on heterogeneous platforms compete for CPU and network resources. In this paper we consider the problem of fairly and efficiently scheduling Bags of Tasks applications on a distributed network of processors organized as a tree. The goal of scheduling is to maximize throughput of each application while ensuring a fair sharing of resources between applications. We particularly investigate decentralized schedulers that use only local information at each participating resource and we assess their performance via simulation, and compare to an optimal centralized solution obtained via linear programming.
Distributed architectures, Scheduling and task partitioning, Linear programming
Olivier Beaumont, Larry Carter, Jeanne Ferrante, Arnaud Legrand, Loris Marchal, Yves Robert, "Centralized versus Distributed Schedulers for Bag-of-Tasks Applications", IEEE Transactions on Parallel & Distributed Systems, vol.19, no. 5, pp. 698-709, May 2008, doi:10.1109/TPDS.2007.70747