The SciMantic project aims at designing and developing a modular software infrastructure termed Semantic Content Infrastructure (SCI). SCI will leverage the Semantic Web technology by providing applications, which are built on top of it, the capability to efficiently relate and search for semantic contents distributed over the Internet and shared securely among independent instances of these applications. The key technologies behind this capability are an integration of Peer-to-Peer (P2P) mechanisms with the Semantic Web.
In SciMantic, a scenario is defined, in which knowledge is shared among community members who have the same interest. Community members can join (and leave) Knowledge Sharing Networks of various topics, e.g., Information Technology, Solar System, and Badminton. Each participant is expected to share his/her knowledge by entering it into the system's database using a sofware tool termed Knowledge Sharing System (KSS). The whole knowledge in the database is divided into (knowledge) units. The participant who enters a particular unit into the database is the unit owner and has the right to manipulate it. Once a unit is stored into the database, all other participants have read access to it, until it is removed from the database. To allow participants to search knowledge of a specific topic, units are tagged with keywords by their owner.
A KSS consists of three services deployed on each KSS Node:
The three KSS services may interact with each other and do not compose any hierarchy. Moreover, any of these services is allowed to interact with users (e.g., via Web browsers).
The key goal of this thesis is a fine design of a P2P-based KSS servicesí architecture, a prototypical implementation of selected services and their evaluation. The results of this thesis have to be able to show the advantages or disadvantages of the combination of Semantic Web technologies with P2P mechanisms in developing a system providing knowledge sharing through Web services running on each Peer. Furthermore, a scalability evaluation of the developed system with respect to the number of Peers as well as the knowledge units shall justify the application of P2P approaches. Finally, the evaluation shall also show whether the combination of Web services, P2P mechanisms, and RDF graph data model provides sufficient performance to share and search distributed knowledge units over the Internet.
Supervisors: Prof. Dr. Burkhard Stillerback to the main page