• DocumentCode
    2708217
  • Title

    Scalable Semantic Overlay Routing in MANETs

  • Author

    Gaddah, Abdulbaset ; Kunz, Thomas ; Kouhi, Ross

  • Author_Institution
    Syst. & Comput. Eng., Carleton Univ., Ottawa, ON, Canada
  • fYear
    2010
  • fDate
    11-14 May 2010
  • Firstpage
    396
  • Lastpage
    402
  • Abstract
    MANETs are an interesting domain for P2P technologies, in particular in the form of content-based routing and semantic overlays. In such an overlay, XML routers match XPATH queries against data in the form of XML documents. Due to the matching complexity, the scalability of such overlays becomes a key concern. Typically, researchers propose to aggregate the queries to reduce the filtering burden on the XML routers. As new XPath queries are added, existing queries deleted, or the topology between routers changes, these aggregate data structures need to be updated in some or all intermediate XML routers. Such updates can be fairly complex, to ensure that previously aggregated queries are now properly matching the new set of queries. We review a number of aggregation approaches and evaluate their scalability in light of a dynamically changing set of queries. Our survey shows that very few aggregation schemes support efficient dynamic updates of the subscription set, in particular in the case of deleting queries. We then show, through a small implementation, how one particular aggregation approach allows updating the aggregate data structure efficiently, while maintaining a high routing performance.
  • Keywords
    Aggregates; Communication networks; Data structures; Filtering; Filters; Routing; Scalability; Subscriptions; Topology; XML; Bloom Filter; MANET; XML; XPATH; content-based overlay; semantic routing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Communication Networks and Services Research Conference (CNSR), 2010 Eighth Annual
  • Conference_Location
    Montreal, QC, Canada
  • Print_ISBN
    978-1-4244-6248-3
  • Type

    conf

  • DOI
    10.1109/CNSR.2010.31
  • Filename
    5489420