eCommons

 

Merging Partitioned Databases

Other Titles

Abstract

Partitioning of a distributed data base requires either that update activity be restricted or that a strategy for conflict resolution and partition merging be used once communication is restored. The graph-theoretic approach used by Davidson follows the latter approach and can be used to show that finding an optimum solution to the general problem is NP-complete. We give several methods of reducing the size of the graphs involved. Two open subproblems are shown to be NP-complete, while an extension of a known polynomial-time subproblem is given. Simulation results are used to study both the amount of compression achieved by the graph reduction techniques and their effects on heuristics for the problem. In addition, some modifications are made to existing heuristics to improve their performance. A probabilistic model is presented and compared with the simulations.

Journal / Series

Volume & Issue

Description

Sponsorship

Date Issued

1983-04

Publisher

Cornell University

Keywords

computer science; technical report

Location

Effective Date

Expiration Date

Sector

Employer

Union

Union Local

NAICS

Number of Workers

Committee Chair

Committee Co-Chair

Committee Member

Degree Discipline

Degree Name

Degree Level

Related Version

Related DOI

Related To

Related Part

Based on Related Item

Has Other Format(s)

Part of Related Item

Related To

Related Publication(s)

Link(s) to Related Publication(s)

References

Link(s) to Reference(s)

Previously Published As

http://techreports.library.cornell.edu:8081/Dienst/UI/1.0/Display/cul.cs/TR83-547

Government Document

ISBN

ISMN

ISSN

Other Identifiers

Rights

Rights URI

Types

technical report

Accessibility Feature

Accessibility Hazard

Accessibility Summary

Link(s) to Catalog Record