Design and analysis of opportunistic forwarding in challenged networks

Challenged networks are those networks, e.g., mobile opportunistic networks (MONs), a.k.a., delay/disruption tolerant networks, and power-constrained or duty-cycled wireless sensor networks (WSNs), where traditional Internet architectures fail to ensure end-to-end communication due to the lack of ‘always-on’ well-connected infrastructures and its resulting intermittent connectivity. Opportunistic forwarding has emerged as a new communication principle relying on node mobility (or relaying information upon contacts between mobile nodes by chance) for effective communication in MONs, while a different form of opportunistic forwarding, or randomized routing, has been also popular for many applications in WSNs due to their desirable properties. In this dissertation, we study the design and analysis of opportunistic forwarding in such challenged networks. Since the challenged networks are highly heterogeneous and dynamic in many aspects, in our study, we carefully take into account, through systematic stochastic analysis, the random, dynamic underlying heterogeneity so as to correctly understand the system behaviors and design new algorithms/procotols to adapt to and exploit the heterogeneity. In the first part of this dissertation, we present our study on analyzing and improving forwarding performance under heterogeneous contact dynamics in MONs. We first discuss how the heterogeneity in mobile nodes’ contact dynamics impacts the forwarding performance in MONs. In particular, we show, through formal stochastic comparisons, that different heterogeneous structures lead to an entirely opposite delay performance, cautioning that one should carefully evaluate the performance of forwarding algorithms under a properly chosen heterogeneous network setting. We next undertake to develop an analytical framework in order to quantify the performance gain achievable by exploiting the heterogeneous contact dynamics to our advantage. The framework enables us to obtain a heterogeneity-aware forwarding policy with its guaranteed delay bound and thus provides quantitative results on the benefit of leveraging underlying heterogeneity structure in the design of forwarding algorithms. In the second part of this dissertation, we study the design of smart/distributed duty-cycling for opportunistic forwarding in heterogeneous and dynamic WSNs toward faster information delivery and longer network lifetime. We first propose and analyze a simple yet effective modification of random duty cycling, named Smart Sleep, for opportunistic forwarding in duty-cycled WSNs. By judiciously exploiting temporal dynamics or intentionally correlating duty-cycling with packet transmission activity, our proposed Smart Sleep breaks the typical delay-power tradeoff and achieves smaller delay as well as more power-saving at each sensor leading to longer network life. We next introduce a distributed wake-up rate control scheme taking advantage of local heterogeneity structure, which is complementary to Smart Sleep, and demonstrate that it also improves both delay and network lifetime for opportunistic forwarding in duty-cycled WSNs.