| Context awareness is the ability to use contextual information in support of operations and decision making. The context dissemination process, an important component to enable context awareness, is essential for the smooth operation of various demanding and self-organizing context-based systems in Ambient Networks. Ambient Networks enable cooperation between heterogeneous networks, on demand, in order to provide seamless connectivity to end-users. We argue that incorporating contexts in a scalable and flexible fashion is a challenging task as it poses overhead and complexity due to non-automated service convergence, and is more difficult to maintain and interoperate successfully. We deal with this challenge and propose a meliorated context dissemination system.;We tend our work towards an overlay design, which acts as the key to assemble and interconnect context nodes and hide the heterogeneity and variability of the underlying networks. We propose a context-specific Pure Overlay Space construction and maintenance architecture on top of overlays to disseminate context information in Ambient Networks with respect to ContextWare (a model for context processing which acquires and distributes context data among interested context entities). We propose a Context Dissemination Processor for processing the Pure Overlay Space's contextual nodes by handling heterogeneous contexts and their respective profiles as agents to attain network interoperability. With the help of the Context Dissemination Processor, the Pure Overlay Space will serve as a middleware and multicast base for the development and maintenance of an application-layer context-aware dissemination protocol, the CSON-D protocol. The overlay infrastructure supports spatial decoupling, semantic clustering, personalization, and agent-based optimizations for virtual customized interaction between context entities. We design and implement a context-specific overlay architecture made up of multi-level overlay networks. The CSON-D protocol's multi-level overlay scheme exhibits adaptive behaviour with minimized casting functionality to disseminate contexts. We also propose an efficient context monitoring scheme with adaptive routing and learning engine to aid the dissemination process. Alongside the conceptual discussion, this thesis is complemented with implementation details with respect to ContextWare and result evaluation. |
