Semantic Tuple Spaces: A Coordination
January 26, 2004
The vision of the Semantic Web expands on the vision of the WWW by associating accessible formal semantics with content and services, and where web pages, databases, personal devices and home appliances are all producers and consumers of information. In a Pervasive Computing environment the user interacts with his environment as opposed to a specific device. Important research challenges in such an environment include auto-configuration of entities, context-sensitive behavior and the creation of unobtrusive services. The Semantic Web can help realize these challenges by providing "semantic interoperability". To enable semantic interoperability, there is a strong need to build "semantic infrastructures" and "semantic gadgets" that would allow heterogeneous entities to work together in a pervasive environment. This thesis presents a Tuple Space based semantic infrastructure that integrates with Vigil, a framework for intelligent services in mobile environments.
Tuple Spaces offer a coordination infrastructure for communication between autonomous entities by providing data persistence, transactional security as well as temporal and referential decoupling-- properties that make it desirable middleware for e-commerce and mobile computing applications. In most Tuple Space implementations tuples are retrieved by employing type-value matching of ordered tuples, object-based polymorphic matching, or XML-based matching. In an heterogeneous environment, interaction between autonomous entities can be enhanced if the retrieval of tuples is based on semantics rather than syntax. This thesis discusses the architecture and implementation of a prototype system, which annotates tuples with semantic description and uses description-logic based reasoning. Specialized agents, like the Tuple-Recommender Agent, Task-Execution Agent and Pubish-Subscribe Agent, that have a better understanding of the environment, reside on the Tuple Space and play an important role in providing user-centric reasoning.
University of Maryland, Baltimore County
Department of computer science and electrical engineering