Pervasive computing is a vision involving future generation computing environments in which information and communication technologies are seamlessly integrated to support diverse aspects of users' lifestyles. Pervasive computing environments will see sophisticated sensors, actuators, and monitoring devices embedded into living environments and interconnected with a variety of computing devices and networking technologies. In the last decade, significant technological advances have led to the proliferation of new devices, such as handheld computers (e.g., PDAs and mobile phones), wearable computers and intelligent home and office appliances. Advances in networking technologies have also seen a shift from wired to wireless networks and increasing availability of wireless access in businesses and academic institutions, as well as in public hot spots located in libraries, coffee shops, and other public places. At the same time, developments in wireless technologies have equipped devices with short range connection capabilities, allowing computing devices and smart appliances located within close range of one another to form micro networks without any central network infrastructure. Current cellular mobile technology is also providing low speed data connections and laying the foundations for future mobile networks that will provide multimedia streaming capabilities to mobile devices.
Together, these emerging technologies are enabling the vision of pervasive computing to take shape, so that users are provided with the control and flexibility to perform their computing tasks in a manner that suits their lifestyles and requirements. Since pervasive computing environments combine heterogeneous networking and device technologies, pervasive systems are required to adapt to changes in the context in which computational applications operate. This adaptation can take various forms, such as adaptation of applications, communication streams, or underlying communication protocols. One possible type of communication protocol adaptation is vertical handover between different networks (i.e., redirection of applications' communication streams from one network to another).
Vertical handover permits users to move freely between heterogeneous networks or change computing devices while maintaining application continuity. However, current vertical handover solutions are not context-aware and are restricted to reactions to disconnections as users move out of network coverage. Since the vertical handover mechanism relies on reactions to disconnections, disruption of the communication stream always occurs during the handover process, and the QoS requirements of the application are likely to be violated. This is often inappropriate, especially for multimedia applications, which have stringent QoS requirements. A better solution for pervasive systems is to perform vertical handovers in response to a wider set of context changes, which may include (i) prediction of users moving in or out of network coverage, (ii) changes in the network, leading to QoS that is unacceptable for applications, (iii) movement of users between different computing devices while continuing their applications, or (iv) users entering preferred networks. This thesis presents an innovative solution for vertical handovers in pervasive systems that can support all of these types of context-triggered handover.
The thesis makes a number of significant research contributions. Firstly, it presents a context model that defines types of context information and context changes that are relevant to vertical handover in pervasive computing environments. Secondly, it develops a set of context evaluation and decision rules for triggering vertical handover and performing stream adaptations to address QoS requirements. Thirdly, it defines a generic vertical handover protocol and architecture for handover between networks or devices. It also presents refinements to the generic solution that take into account features of particular networking technologies. Fourthly, the thesis illustrates mechanisms that can be incorporated into the vertical handover solution to counter and minimise QoS violations during changes between networks, in order to minimise application disruption and maximise handover transparency for the user. Finally, as a proof of concept, the thesis presents a prototype that demonstrates the ability of the proposed vertical handover architecture to support seamless computing, and also enables performance evaluation for various types of networking technology.