| Abstract |
An increasing interest in Community Web Portals (corporate, e-marketplace etc.) can be observed over the last few years. The core Portal component is the Knowledge Catalog holding different types of metadata(e.g. cataloguing, content rating,) associated with different types of portal resources (e.g documents, data) that are available to the community members. The Knowledge Catalog supports the organization of the content of the portal in a multitude of ways by employing ontologies and theusari. Fundamental services of a Portal, such as queryning in semantically meaningful way, personalization and integration of heterogeneous sources, are based on the knowledge Catalog. The expected large volume of the knowledge catalog(for instance, the Open Directory Portal of Netscape comprisrs around 180M of Subject Topics and 700M of indexed URLS) makes the management of metadata using database management systems a challenging topic. RDF, a W3C recommendation is by now a standard framework for the description of resources. The topic of this thesis is the representation and the efficient storage of RDF metadata in object-relational database management systems (ORDBMS). Due to the peculiarities of RDF in comparison to conventional data model (e.g. properties are first class objects, optional and multivalued properties, specialization between properties), its representation in ORDBMS is fairly complex. In order to define the semantic constraints which metadata of a community should conform to and to achieve its effective storage, we study the theoretical underpinnings of the problem using the knowledge representation language Telos. Our general model is expressive enough to represent the majority of RDF primitives as well as multiple instantiation. It also considers tha RDF graphs have labels on both nodes and edges, optimizes processing of class/property hierarchies and ensures better query performance. A distinctive feature of our system is the fact that it enables incremental loading of RDF desriptions and schemas, wich is crucial for handling tha large volumes of RDF metadata.
|
Αποθήκευση Μεταδεδομένων RDF για Πύλες Κοινοτήτων Διαδικτύου
