Optimal sharing of bags of tasks in heterogeneous clusters
- 7 June 2003
- proceedings article
- Published by Association for Computing Machinery (ACM)
Abstract
We prove that "FIFO" worksharing protocols provide asymptotically optimal solutions to a problem related to sharing a bag of identically complex tasks in a heterogeneous network of workstations (HNOW) n. In the HNOW-Exploitation Problem, one seeks to accomplish as much work as possible on n during a prespecified fixed period of L time units. The worksharing protocols we study are crafted within an architectural model that characterizes n via parameters that measure workstations' computational and communicational powers. The protocols are self-scheduling, in that they determine completely both an amount of work to allocate to each of n's workstations and a schedule for all related interworkstation communications. A protocol observes a FIFO regimen if it has n's workstations finish their assigned work, and return their results, in the same order in which they are supplied with their workloads. The optimality of FIFO protocols resides in the fact that they accomplish at least as much work as any other protocol during all sufficiently long worksharing episodes. Simulation experiments indicate that the superiority of FIFO protocols is often observed during worksharing episodes of only a few minutes' duration.Keywords
This publication has 12 references indexed in Scilit:
- Broadcast scheduling optimization for heterogeneous cluster systemsPublished by Association for Computing Machinery (ACM) ,2000
- Task allocation on a network of processorsIEEE Transactions on Computers, 2000
- Designing communication strategies for heterogeneous parallel systemsParallel Computing, 1998
- LogPCommunications of the ACM, 1996
- LogGPPublished by Association for Computing Machinery (ACM) ,1995
- A case for NOW (Networks of Workstations)IEEE Micro, 1995
- Optimal sequencing and arrangement in distributed single-level tree networks with communication delaysIEEE Transactions on Parallel and Distributed Systems, 1994
- Optimal broadcast and summation in the LogP modelPublished by Association for Computing Machinery (ACM) ,1993
- Models and algorithms for coscheduling compute-intensive taks on a network of workstationsJournal of Parallel and Distributed Computing, 1992
- Distributed computation for a tree network with communication delaysIEEE Transactions on Aerospace and Electronic Systems, 1990