A study of meta-scheduling architectures for high throughput computing

Abstract : In this paper we present a model and a simulator for large-scale system. Such platforms are composed of heterogeneous clusters of PCs belonging to a local network. These clusters are then connected to each other through a global network. Moreover each cluster is managed via a local scheduler and is shared by many users. We validate our simulator by comparing the experimental results and the analytical results of a M/M/4 queuing system. These studies indicate that the simulator is consistent. After that we do the comparison with a real batch system and we obtain a mean error of 10.5 \% for the response time and 12 \% for the makespan. We conclude that our simulator is realistic and describes well the behavior of a large-scale system. Thus we can study the scheduling of our system called \dirac in a high throughput context. We justify our decentralized, adaptive and opportunistic approach in comparison to a centralized approach in such a context.
Document type :
Reports
Complete list of metadatas

Cited literature [18 references]  Display  Hide  Download

https://hal-lara.archives-ouvertes.fr/hal-02101915
Contributor : Colette Orange <>
Submitted on : Wednesday, April 17, 2019 - 9:09:41 AM
Last modification on : Friday, May 17, 2019 - 1:39:22 AM

File

RR2005-13.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02101915, version 1

Collections

Citation

Eddy Caron, Vincent Garonne, Andreï Tsaregorodtsev. A study of meta-scheduling architectures for high throughput computing. [Research Report] LIP RR-2005-13, Laboratoire de l'informatique du parallélisme. 2005, 2+13p. ⟨hal-02101915⟩

Share

Metrics

Record views

2

Files downloads

7