A failure detector for HPC platforms

George Bosilca, Aurelien Bouteiller, Amina Guermouche, Thomas Herault, Yves Robert, Pierre Sens, Jack Dongarra

Research output: Contribution to journalArticlepeer-review

15 Scopus citations

Abstract

Building an infrastructure for exascale applications requires, in addition to many other key components, a stable and efficient failure detector. This article describes the design and evaluation of a robust failure detector that can maintain and distribute the correct list of alive resources within proven and scalable bounds. The detection and distribution of the fault information follow different overlay topologies that together guarantee minimal disturbance to the applications. A virtual observation ring minimizes the overhead by allowing each node to be observed by another single node, providing an unobtrusive behavior. The propagation stage uses a nonuniform variant of a reliable broadcast over a circulant graph overlay network and guarantees a logarithmic fault propagation. Extensive simulations, together with experiments on the Titan Oak Ridge National Laboratory supercomputer, show that the algorithm performs extremely well and exhibits all the desired properties of an exascale-ready algorithm.

Original languageEnglish
Pages (from-to)139-158
Number of pages20
JournalInternational Journal of High Performance Computing Applications
Volume32
Issue number1
DOIs
StatePublished - Jan 1 2018

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is partially supported by the CREST project of the Japan Science and Technology Agency (JST), by NSF grant #1339820, by the PIA ELCI (Bull Inria) project and partially supported by the NSF (award \#1564133).

Keywords

  • MPI
  • failure detection
  • fault tolerance

Fingerprint

Dive into the research topics of 'A failure detector for HPC platforms'. Together they form a unique fingerprint.

Cite this