Context Middleware

Policy-Based Architecture for Context Management In Ambient Networks:

The ultimate vision of ambient networking is to create a unified and transparent view of the otherwise heterogeneous networking environments. In such environments, computing devices and offered services will be capable of intuitively and proactively serve users in their everyday activities in an invisible and unobtrusive manner. While advances in both sensory devices and communication technologies represent a step forward towards the realization of the true essence of ambient networking, a major challenge remains that is of making devices and applications aware of changes in their surrounding environments to best adjust their functionalities and behavior in a non intrusive manner. Context-awareness is an emerging field of research that aims at providing applications with the means to discover and reason about a situation and adapt to it in a dynamically changing environment.

Earlier attempts to model and utilize context were mainly user-centric and focused only on a predefined set of information (e.g., users’ location and identity). In these approaches, attribute-value pairs provided a simple and sufficient representation of context. With the emergence of new types of ubiquitous applications, context took a new definition to include all information related to any entity in the environment (e.g., humans, devices, software agents and components of virtual environments). To this end, the development of a flexible yet powerful context model that can cover a wide range of possible contexts became a challenging task.

This project introduces the notion of context-policies as a vehicle for representing and disseminating context. The aim is to develop architectures for a policy-centered middleware providing policy structures to help describe and control the multifacets of an entity’s context. In addition to disseminating context, the utilization of context-policies in allows the middleware to make decisions on behalf of some context clients by directly selecting and executing a preexisting policy action if a situation occurs in a given context, or by assembling context-policies dynamically so that services and applications can behave according to the new context. Moreover, context processing and management operations, performed through enforcement of different policies, should not impose any computational limitation as compared to other logic based models.