Dynamic Co-Scheduling of Distributed Computation and Replication

Authors:
Huadong Liu;Micah Beck;Jian Huang
Affiliations:
The University of Tennessee, USA;The University of Tennessee, USA;University of Tennessee, USA
Venue:
CCGRID '06 Proceedings of the Sixth IEEE International Symposium on Cluster Computing and the Grid
Year:
2006

Citing 0
Cited 4

Batch Coloring Flat Graphs and Thin

SWAT '08 Proceedings of the 11th Scandinavian workshop on Algorithm Theory
DECO: data replication and execution CO-scheduling for utility grids

ICSOC'06 Proceedings of the 4th international conference on Service-Oriented Computing
Max-coloring paths: tight bounds and extensions

Journal of Combinatorial Optimization
Time-critical distributed visualization with fault tolerance

EG PGV'08 Proceedings of the 8th Eurographics conference on Parallel Graphics and Visualization

Quantified Score

Hi-index	0.00

Visualization

Abstract

We are interested in developing the infrastructural tools that allow a distributed data intensive computing environment to be shared by a group of collaborating but geographically separated researchers in an interactive manner, as opposed to a batch mode of operation. However, without advanced reservation, it is difficult to assure a certain level of performance on a large number of shared and heterogeneous servers. To achieve scalable parallel speedups in this scenario, we must closely integrate the management of computation and runtime data movement. In this paper, we first define the canonical scheduling problem for datasets distributed with k-way replication in the wide area. We then develop a dynamic coscheduling algorithm that integrates the scheduling of computation and data movement. Using time-varying visualization as the driving application, we demonstrate that our co-scheduling approach improves not only application performance but also server utilization at a very reasonable cost.