UMBC ebiquity
CoCoNet: Content and Context Aware Networking
July 1, 2005 - December 1, 2007
The current Internet was originally designed to provide best-effort data transport over a wired infrastructure with end hosts utilizing a layered network stack to provide reliability, quality of service, security etc. for user applications. However, the proliferation of inelastic applications, coupled with wide spread migration towards hybrid networks utilizing wired and wireless links and the plethora of end host variants ranging from cell phones to enterprise servers necessitates the migration of more and more services away from the edges and into the network. The aim of this thesis is to provide a generic and flexible framework and associated algorithms that will enable the incremental deployment of intelligent services into the network with the aim of optimizing the end-user experience for networked applications. We focus our research on two key facets; cross layer optimization algorithms to enable efficient transport of data across a hybrid network coupled with the inclusion of semantic information in the data packets that can be intelligently processed within the network. The aim is to show that if routers in a network have visibility into the type of data that they are currently handling (either at a packet level or a flow level), the routers can then perform optimizations and content adaptations relevant to that specific type of data based on local policies. We intend to use RDF/RDFS as the medium to convey this semantic information thereby allowing interim routers to reason over their existing knowledge base on how to specifically handle a given data packet or stream in a flexible and generic manner. Mechanisms to incrementally deploy our framework into a large scale network along with the implementation of new value added services that can now be offered will be demonstrated using a large scale testbed.