Exploring intelligent service migration in a highly mobile network

Mobile services allow services to be migrated or replicated closer to users as they move around. This is now regarded as a viable mechanism to provide good Quality of Service to users in highly mobile environments such as vehicular networks. The vehicular environment is rapidly becoming a significant part of the internet and this presents various challenges that must be addressed; this is due to continuous handovers as mobile devices change their point of attachment to these networks resulting in a loss of service. Therefore, this explains the need to build a framework for intelligent service migration. This thesis addresses these issues.
It starts by discussing the requirements for intelligent service migration. Then it investigates a low latency Quality of Service Aware Framework as well as an experimental transport protocol that would be favoured by vehicular networks.
Furthermore, two analytical models are developed using the Zero-Server Markov Chain technique which is a way of analysing scenarios when the server is not continuously available to serve. Using the Zero-Server Markov Chain, the first analytical model looks at lost service due to continuous handovers and the communication dynamics of vehicular networks, while the second model analyses how service migration affects service delivery in these networks. Formulas are developed to yield the average number of packets in the system, the response time, the probability of blocking and a new parameter called the probability of lost service. These formulas are then applied to the Middlesex VANET Testbed to look at reactive and proactive service migration. These techniques are then incorporated into a new Service Management Framework to provide sustainable Quality of Service and Quality of Experience to mobile users in vehicular networks. This thesis also shows that this new approach is better than current approaches as it addresses key issues in intelligent service migration in such environments, and hence can play a significant part in the development of Intelligent Transport Systems for Smart Cities.