Scott M. Pike

Department of Computer Science, Texas A&M University

Department of Computer Science, Texas A&M University
  • home »
  • publications »
  • View all publications
back

Links

  • Home
  • +  Research
    • Main research page
    • Research Project 1
    • Research Project 2
  • +  Teaching
    • Main teaching page
    • CPSC-668
    • CPSC-689
    • CPSC-410/611
  • +  Publications
    • All publications
    • Book Chapters
    • Conference Proceedings
    • Workshop Papers
  • Service
  • Resources
  • Personal
  • Current CV

Contact

  • pike[at]cse.tamu.edu
  • +1.979.776.2162 (tel)
  • +1.979.847.8578 (fax)
13

Publications

Publications listed by category and date of publication.

Please observe the copyright notices posted here governing the use of material from this page.

Click for papers indexed by

ACM, IEEE, DBLP, Springer, CiteSeer, and Google Scholar

Sort:

By Date: Newest publications first | Oldest publications first
By Category: Book Chapters | Conference Proceedings | Workshop Papers
By Publisher: ACM | ACTA | Cambridge | ETD | IEEE Xplore | None | Springer



ACTA


  A New Distributed Resource-Allocation Algorithm with Optimal Failure Locality

  2000 [more]
ACTAs Icon
  • Citation:
    Paolo A.G. Sivilotti, Scott M. Pike, and Nigamanth Sridhar, "A New Distributed Resource-Allocation Algorithm with Optimal Failure Locality." in Proceedings of PDCS 2000, Vol 2, pp. 524-529, IASTED/ACTA Press, 2000. Citations: [1] [2] [3] [4]
  • Abstract:
    Failure locality measures an algorithm's robustness to process failures. We present a new algorithm for the dining philosophers problem | a classic problem in distributed resource allocation | that has optimal failure locality. As a re nement, the algorithm can be easily parameterized by a simple failure model to achieve super-optimal failure locality in the average case. Keywords: distributed algorithms, fault tolerance, dining philosophers, failure locality. 1 Introduction The dining philosophers problem is a classic and fundamental resource allocation problem [6]. Although rst formulated as a shared-memory concurrency problem, it has since received considerable attention as a distributed conict-resolution problem [8]. It can be seen as a generalization of the mutual exclusion problem, in which neighboring processes cannot access a shared resource simultaneously. It has many applications in the construction of other distributed resourceallocation algorithms, including drinking p...
  • Publisher: ACTA
  • Link to copy of this pubilcation on file with the publisher:
    www.actapress.com/proceedings/pdcs.htm

  • Download this publication:  
    faultloc.pdf faultloc.pdf
    faultloc.ps faultloc.ps
    PDCS2000.bib PDCS2000.bib




Total Number of Publications: 13
publication admin

Last Modified: Mon Jul 21 11:14:57 CDT 2008
© Scott M. Pike (some rights reserved...)     My Erdös Number | Copyright | Site Map | Contact | About