Distributed wireless channel allocation in networks with mobile base stations

In traditional cellular networks with fixed base stations, the channel reuse pattern is static and deterministic. When the cell layout is dynamic, due to the mobility of base stations, the cluster of cells within cochannel interference range changes with time. Consequently, the channel reuse pattern is highly dynamic. Moreover, base stations also need wireless channels to communicate amongst themselves. A communication session between a pair of nodes may have to switch channels due to the movement of other nodes into the neighborhood. None of the existing dynamic channel allocation algorithms for cellular networks works in such a system. Hence, there is a need for new wireless channel allocation algorithms for virtual cellular networks with mobile base stations. In this paper, principles of mutual exclusion pertaining to distributed computing systems are employed to develop such an algorithm. The inter-base-station wireless links are referred to as backbone links, while the base station to mobile node links are referred to as short-hop links. The proposed algorithm is distributed, dynamic, and deadlock-free. Disjoint sets of channels are used for backbone and short-hop links. The distributed nature of the channel allocation scheme leads to scalability and robustness, as the responsibility is no longer centralized at the mobile telecommunications switching office (MTSO). Instead, it is shared among all the mobile base stations. In addition, the issue of channel rearrangement is addressed. Channel rearrangement is the switching of channels, performed to prevent cochannel interference, when mobile base stations using the same channels, hitherto not in interference range, come within the range of each other. If multiple channels are available to support a communication session, the channel selection policy can have a significant impact on performance. A random selection from the set of available channels yields better performance at low to moderate channel demand. This is in contrast to ordered channel selection (from one end of the spectrum), which always yields the best performance for cellular networks. Results obtained by simulating the algorithm are consistent with the theoretically obtained values.