Information integration in schema-based peer-to-peer networks

  • Authors:

  • Alexander Löser;Wolf Siberski;Martin Wolpers;Wolfgang Nejdl

  • Affiliations:

  • CIS, Technical University Berlin, Berlin, Germany; ; ;Learning Lab Lower Saxony, University of Hannover, Hannover, Germany

  • Venue:
  • CAiSE'03 Proceedings of the 15th international conference on Advanced information systems engineering
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Peer-to-peer (P2P) networks have become an important infrastructure during the last years. Using P2P networks for distributed information systems allows us to shift the focus from centrally organized to distributed information systems where all peers can provide and have access to information. In previous papers, we have described an RDF-based P2P infrastructure called Edutella which is a specific example of a more advanced approach to P2P networks called schema-based peer-to-peer networks. Schema-based P2P networks have a number of advantages compared with simpler P2P networks such as Napster or Gnutella. Instead of prescribing one global schema to describe content, they support arbitrary metadata schemas and ontologies (crucial for the Semantic Web). Thereby they allow complex and extendable descriptions of resources thus introducing dynamic behavior to the former fixed and limited descriptions, and can provide complex query facilities against these metadata instead of simple keyword-based searches. In this paper we will elaborate topologies, indices and query routing strategies for efficient query distribution in such networks. Our work is based on the concept of super-peer networks which provide better scalability compared to traditional P2P networks. By adapting existing concepts of mediator-based information systems to super-peer based networks, as we will showin this paper, they are able to support sophisticated routing, clustering and mediation strategies based on the metadata schemas and attributes. The resulting routing indices can be built using local clustering policies and support local mediation and transformation rules between heterogeneous schemas, and we sketch some first ideas for implementing these advanced functionalities as well.