Towards Entity-Centric Wide-Area Context Discovery

Mobile computing leads to different usage situations for the same service due to the mobility of the user. Context aware services (CASs) try to exploit information about the current usage situation in order to better satisfy the user's needs. Especially on a large-scale, CASs face the problem of discovering appropriate context information services (CISs), which are capable of delivering a particular type of context information about a particular entity. The mapping between CISs and entities, however, is highly dynamic, especially when entities are mobile and CISs have only limited coverage areas. New mechanisms are needed for organizing CISs and their currently covered entities in a scalable, robust and decentralized way so that they can be discovered by (potentially mobile) CASs when required. As a possible solution we present the design of an entity-centric wide- area context discovery service based on distributed hashtables. It requires no dedicated infrastructure, and ontologies are used for achieving interoperability. The concept is evaluated by a prototype based on the OpenDHT infrastructure. Experimental results demonstrate the feasibility and performance of the approach.