Research On A Few Important Problems In Wireless Networks

This dissertation investigates a few important challenging problems, such as network capacity, the application of network coding and network security, in wireless ad hoc networks (WANETs).First, we study the throughput capacity and delay for wireless ad hoc networks with partial infrastructure. It has been shown that only when the number of base stations m and the number of normal nodes are on the same order, can the throughput capacity be linearly increasing with the number of nodes (the scaling law). Moreover, through the use of newly emerging MIMO distributed transmissions schemes, we can also achieve the scaling law for pure wireless ad hoc networks. One question remains: for mixed wireless ad hoc networks with partial infrastructure, can we use MIMO technologies to make the throughput capacity scale? We investigate this problem and point out that with the current resource allocation strategy commonly used today, we cannot make the throughput capacity scale for hybrid ad hoc networks. We have shown that the hybrid WANETs can scale only when one of the following two conditions holds: (1) when m =Θ(1 ) and all resource is allocated to the ad hoc transmission mode; (2) when m =Θ( n) and all resource is allocated to infrastructure transmission mode. Moreover, the delay is very large for the former while isΘ(1 ) for the latter.Second, we look into the application of network coding in wireless ad hoc networks and propose a novel practical network coding scheme to deal with intra-flow and inter-flow traffics. This scheme is based on the previously known COPE and allows each node to use intra-flow network coding to enhance the transmission reliability in order to increase the throughput capacity. Based on a simplified network topology, we derive some analytical results and theoretically demonstrate the feasibility and effectiveness of our proposed scheme. We also propose a multipath transmission scheme to enhance the reliability between two nodes for WANETs with low transmission reliability, and further increase the throughput capacity. We discuss how to effectively utilize network coding and objectively assess the advantages and disadvantages of our scheme.Finally, we investigate the wireless network security. In WANETs, due to the openness of wireless routing paths, the lack of existing infrastructure, and the limited computational and communication capability, many security schemes for wireline networks or traditional wireless networks may not be effective. Moreover, most WANETs are deployed for sensitive applications with high demand on security, and hence it is vital to search for high efficient and feasible security mechanism. In our final research task, we investigate the security problem for resource-constrained WANETs. We show that ID-based cryptography is perfect fit for addressing the security issues in resource-constrained WANETs, and then apply this technique to wireless sensor networks and wireless mesh networks for authentication and key establishment, and come up with a location-based security scheme for wireless sensor networks and a simple yet effective secure access scheme for wireless mesh networks. By observing that wireless channels are lossy and the receiving ends are typically resource-limited for multicast services, we propose a novel authentication scheme MABS, which not only mitigates the computational burden in the signature verifications, but also is resilient to packet loss. Moreover, its enhanced version MABS-E can also effectively counteract the denial of service attack.