Automated Analysis of Impact of Scheduling on Performance of Self-Stabilizing Protocols

Abstract : In a concurrent computing system, a scheduler determines at each time which computing task should execute next. Thus, a scheduler has tremendous impact on the performance of the tasks that it orchestrates. Analyzing the impact of scheduling in a distributed setting is a challenging task, as it is concerned with subtle dimensions such as geographical distance of processes and the achievable level of parallelism. In this paper, we propose an automated method based on probabilistic verification for analyzing fault recovery time in distributed self-stabilizing protocols. We exhibit the usefulness of our approach through a large set of experiments that demonstrate the impact of different types of scheduling policies on recovery time of different classes of stabilizing protocols, and the practical efficiency of classical self-stabilizing scheduler transformers.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01213273
Contributor : Lip6 Publications <>
Submitted on : Thursday, October 8, 2015 - 10:52:43 AM
Last modification on : Tuesday, May 14, 2019 - 10:14:30 AM

Identifiers

Citation

Saba Aflaki, Borzoo Bonakdarpour, Sébastien Tixeuil. Automated Analysis of Impact of Scheduling on Performance of Self-Stabilizing Protocols. The International Conference on Stabilization, Safety, and Security in Distributed Systems (SSS 2015), Aug 2015, Edmonton, Canada. pp.156-170, ⟨10.1007/978-3-319-21741-3_11⟩. ⟨hal-01213273⟩

Share

Metrics

Record views

167