Optimal Throughput–Delay Tradeoff in MANETs With Supportive Infrastructure Using Random Linear Coding

The performance of throughput, delay, and their tradeoff in mobile ad hoc networks (MANETs) has been investigated on different assumptions. Nevertheless, few papers consider the supportive infrastructure in MANETs. With the help of supportive infrastructure, i.e., cellular networks, the performance of throughput and delay in MANETs can be improved. The impact of supportive infrastructure on the throughput-delay tradeoff in MANETs is an open research point. In this paper, we are interested in the optimal throughput-delay tradeoff in MANETs with supportive infrastructure. We investigate the supportive infrastructure that only provides transmission pipes between distant nodes. We study the optimal throughput-delay tradeoff in MANETs on different assumptions such as different densities of nodes and base stations (BSs), discrete/continuous mobility models with various velocities, and the fast/slow mobility assumption. We obtain the asymptotically optimal throughput-delay tradeoff and propose transmission policies based on random linear coding (RLC) to achieve the optimal throughput-delay tradeoff asymptotically. From the obtained optimal throughput-delay tradeoff, we find that supportive infrastructure reduces the delay bound from velocity, which is a lower bound of delay caused by nodes' velocity. In particular, if supportive infrastructure can cover the entire network, then the delay bound from velocity vanishes when per-node throughput does not exceed a threshold. Moreover, we find that there also exists a delay bound from the transmission range of the instant transmission. In addition, we observe that, for the optimal throughput-delay tradeoff in the large throughput region, the case in the continuous mobility model outperforms the case in the discrete mobility model, and the case on the slow mobility assumption outperforms the case on the fast mobility assumption.