Simple and efficient network decomposition and synchronization

Shlomo Moran, Sagi Snir

Research output: Contribution to journalArticlepeer-review

Abstract

We present a simple and efficient method for constructing sparse decompositions of networks. This method is used to construct the sparse decompositions needed for variants of the synchro-nizers in [2, 15] in O(|V|) time and O(|E| + |F| log |V|) communication complexities, while maintaining constant messages size and constant memory per edge. Using these decompositions, we present simple and efficient variants of the synchronizers in the above papers. For example, our constructions enable to perform Breadth First Search in an asynchronous network, in which no preprocessing had been done, in communication and time complexities of O(K|V|D + |E| + |V| log |V|) and O(D logK |V| + |V|), respectively, where K ≥ 2 is a parameter, and D is the diameter of the network. We also present an efficient cover-coarsening algorithm, which uses a novel technique for efficient merging of clusters, and improves previous coarsening algorithms in several aspects.

Original languageEnglish
Pages (from-to)217-241
Number of pages25
JournalTheoretical Computer Science
Volume243
Issue number1-2
DOIs
StatePublished - 28 Jul 2000
Externally publishedYes

Bibliographical note

Funding Information:
∗Supported in part by the Bernard Elkin Chair in Computer Science, and by US{Israel BSF grant 95-00328. E-mail address: moran@cs.technion.ac.il (S. Moran).

Keywords

  • Asynchronous communication networks
  • Graph decompositions
  • Synchronizers

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Simple and efficient network decomposition and synchronization'. Together they form a unique fingerprint.

Cite this