Exploiting user contention to optimize proactive resource allocation in future networks

In order to provide ubiquitous communication, seamless connectivity is now required in all environments including highly mobile networks. By using vertical handover techniques it is possible to provide uninterrupted communication as connections are dynamically switched between wireless networks as users move around. However, in a highly mobile environment, traditional reactive approaches to handover are inadequate. Therefore, proactive handover techniques, in which mobile nodes attempt to determine the best time and place to handover to local networks, are actively being investigated in the context of next generation mobile networks. The Y-Comm Framework which looks at proactive handover techniques has de�fined two key parameters: Time Before Handover and the Network Dwell Time, for any given network topology. Using this approach, it is possible to enhance resource management in common networks using probabilistic mechanisms because it is now possible to express contention for resources in terms of: No Contention, Partial Contention and Full Contention. As network resources are shared between many users, resource management must be a key part of any communication system as it is needed to provide seamless communication and to ensure that applications and servers receive their required Quality-of-Service. In this thesis, the contention for channel resources being allocated to mobile nodes is analysed. The work presents a new methodology to support proactive resource allocation for emerging future networks such as Vehicular Ad-Hoc Networks (VANETs) by allowing us to calculate the probability of contention based on user demand of network resources. These results are veri�ed using simulation. In addition, this proactive approach is further enhanced by the use of a contention queue to detect contention between incoming requests and those waiting for service. This thesis also presents a new methodology to support proactive resource allocation for future networks such as Vehicular Ad-Hoc Networks. The proposed approach has been applied to a vehicular testbed and results are presented that show that this approach can improve overall network performance in mobile heterogeneous environments. The results show that the analysis of user contention does provide a proactive mechanism to improve the performance of resource allocation in mobile networks.