Semantic Matching in the SWAP Project

The current state-of-the-art in Knowledge Management solutions still focuses on one or a relatively small number of highly centralized knowledge repositories with ontologies as the conceptual backbone for knowledge brokering. As it turns out, this assumption is very restrictive, because, (i), it creates major bottlenecks and entails significant administrative overheads, especially when it comes to scaling up to large and complex problems; (ii), it does not lend itself to easy maintenance and the dynamic updates often required to reflect changing user needs, dynamic enterprise processes or new market conditions. In contrast Peer-to-Peer computing (P2P) offers the promise of lifting many of these limitations. At the same time, today’s P2P solutions support only limited update, search and retrieval functionality, e.g. search in Napster is restricted to string matches involving just two fields: “artist” and “track”. These flaws however make current P2P systems unsuitable for knowledge sharing purposes. The SWAP project aims at a P2P based knowledge management system that integrates the advantages of Semantic Web-based knowledge management technology developed in successful IST projects like On-To-Knowledge, KnowNet, or Comma. SWAP aims at benefits of a P2P based system that show just by installing the client software, viz. immediate automatic access to knowledge stored at peers. Of course, explicitly modelled ontologies may increase the benefits brought by any knowledge management solution, because they may improve the accuracy of knowledge access and sharing. SWAP solutions, however, may produce benefits even with near zero investment in contrast to conventional knowledge management systems that need an extensive and expensive set-up phase. Conventional knowledge management repositories will still appear as just another, powerful peer in the network. Hence, a combined Semantic Web and P2P solution may always outperform the sophisticated, but conventional centralized system.