In order for pervasive computing systems to be able to perform tasks which support us in everyday life without requiring attention from the users of the environment, they need to adapt themselves in response to context. This makes context-awareness in general, and context-aware adaptation in particular, an essential requirement for pervasive computing systems. Two of the features of context-awareness are: contextual reconfiguration and contextual adaptation in which applications adapt their behaviour in response to context. We combine both these features of context-awareness to provide a broad scope of adaptation and put forward a system, called Policy-Based Contextual Reconfiguration and Adaptation (PCRA) that provides runtime support for both. The combination of both context-aware reconfiguration and context-aware adaptation provides a broad scope of adaptation and hence allows the development of diverse adaptive context-aware applications. However, another important issue is the choice of an effective means for developing, modifying and extending such applications. The main argument forming the basis of this thesis is that we advocate the use of a policy-based programming model and argue that it provides more effective means for developing, modifying and extending such applications. This thesis addresses other important surrounding issues which are associated with adaptive context-aware applications. These include the management of invalid bindings and the provision of seamless caching support for remote services involved in bindings for improved performance. The bindings may become invalid due to failure conditions that can arise due to network problems or migration of software components, causing bindings between the application component and remote service to become invalid. We have integrated reconfiguration support to manage bindings, and seamless caching support for remote services in PCRA. This thesis also describes the design and implementation of PCRA, which enables development of adaptive context-aware applications using policy specifications. Within PCRA, adaptive context-aware applications are modelled by specifying binding policies and adaptation policies. The use of policies within PCRA simplifies the development task because policies are expressed at a high-level of abstraction, and are expressed independently of each other. PCRA also allows the dynamic modification of applications since policies are independent units of execution and can be dynamically loaded and removed from the system. This is a powerful and useful capability as applications may evolve over time, i.e. the user needs and preferences may change, but re-starting is undesirable. We evaluate PCRA by comparing its features to other systems in the literature, and by performance measures.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:554837 |
| Date | January 2011 |
| Creators | Dhomeja, Lachhman Das |
| Publisher | University of Sussex |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://sro.sussex.ac.uk/id/eprint/6933/ |
Page generated in 0.002 seconds