Dynamic partitioning of IP-based wireless access networks

In HMIPv6-based network architecture, Mobile Nodes (MNs) generate excessive signalling overhead, where they move at borders of different Mobility Anchor Point (MAP) domains. Furthermore all the traffic originated or destined to MNs must traverse through the MAPs. Thus, MAPs are potential points of bottlenecks in the network, as the traffic load grows. To mitigate these drawbacks, division of the access network domain into overlapping partitions (MAP domains) is proposed in order to provide more scalable and robust mobile services. Partitioning problem is an NP-hard problem. In this paper, we propose three, heuristic, Kernighan–Lin based partitioning algorithms to minimise the total inter-area handover rate as well as bottleneck effect of MAPs. Our proposed algorithms adopt the proposed scheme of multiple MAP deployment per domain, where each AR can be assigned to more than one MAP in the same level of hierarchy. The proposed algorithms dynamically adapt to traffic and mobility changes. We devise a cost function for each algorithm and formulate optimisation problems, accordingly. Then we evaluate the performance of the algorithms in terms of dynamic adaptation to mobility and load changes, and also to the degree of load-balance, and mean amount of bandwidth blocking and dropping rates. A simulation evaluation and comparison study with Sanchis algorithm, demonstrates that the proposed overlapping scheme outperforms the conventional non-overlapping scheme. Furthermore, all proposed algorithms obtain improved performance implementation of Sanchis algorithm.