Efficient management of a distributed system is a common problem for university’s and commercial computer centres, and handling node failures is a major aspect of it. Failures which are rare in a small commodity cluster, at large scale become common, and there should be a way to overcome them without restarting all parallel processes of an application. The efficiency of existing methods can be improved by forming a hierarchy of distributed processes. That way only lower levels of the hierarchy need to be restarted in case of a leaf node failure, and only root node needs special treatment. Process hierarchy changes in real time and the workload is dynamically rebalanced across online nodes. This approach makes it possible to implement efficient partial restart of a parallel application, and transactional behaviour for computer centre service tasks.
Original languageEnglish
Title of host publicationComputational Science and Its Applications - ICCSA 2015
Subtitle of host publication15th International Conference, Banff, AB, Canada, June 22-25, 2015, Proceedings, Part IV
PublisherSpringer Nature
Pages259-271
ISBN (Electronic)978-3-319-21410-8
ISBN (Print)978-3-319-21409-2
DOIs
StatePublished - 2015
Event15th International Conference on Computational Science and Its Applications, ICCSA 2015 - Banff, Canada
Duration: 21 Jun 201524 Jun 2015

Publication series

NameLecture Notes in Computer Science
PublisherSpringer Nature
Volume9158
ISSN (Print)0302-9743

Conference

Conference15th International Conference on Computational Science and Its Applications, ICCSA 2015
Country/TerritoryCanada
CityBanff
Period21/06/1524/06/15

    Research areas

  • Long-lived transactions, Distributed pipeline, Node discovery, Software engineering, Distributed computing, Cluster computing

ID: 71354892