Tree structure for expressive MapReduce framework

Nathar Shah, Chris Messom

    Research output: Chapter in Book/Report/Conference proceedingConference PaperResearch

    Abstract

    The traditional MapReduce paradigm lacks in expressivity. However it is popular for its fault-tolerance, automatic parallelism, and the ease of distributed computing. Our work aims to tackle this by introducing an expressive MapReduce framework. Expressivity has its theoretical foundations in description logics. Itallows for reasoning tasks such as subsumption. We have picked description logic EL+ for the expressivity due to its feasible computational complexity. In this work we will show how a tree structure be used as the basis for expressive EL+ MapReduce. Trees will be pruned to sub-trees and re-assembled in parallel and distributed MapReduce framework. As a result, large scale linked data reasoning on MapReduce framework would be possible in the expressive MapReduce framework.
    Original languageEnglish
    Title of host publication2017 International Conference on Software and e-Business (ICSEB 2017)
    Subtitle of host publicationDecember 28-30, 2017 Hong Kong
    EditorsShuanghua Yang
    Place of PublicationNew York NY USA
    PublisherAssociation for Computing Machinery (ACM)
    Pages33-37
    Number of pages5
    ISBN (Electronic)9781450354882
    DOIs
    Publication statusPublished - 2017
    EventInternational Conference on Software and e-Business 2017 - Hong Kong, Hong Kong
    Duration: 28 Dec 201730 Dec 2017
    https://web.archive.org/web/20170524060909/http://www.icseb.org/

    Conference

    ConferenceInternational Conference on Software and e-Business 2017
    Abbreviated titleICSEB 2017
    Country/TerritoryHong Kong
    CityHong Kong
    Period28/12/1730/12/17
    Internet address

    Keywords

    • Big data
    • Description logic
    • MapReduce
    • Tree structure

    Cite this