Resolving the Connectivity-Throughput Trade-Off in Random Networks

TL;DR

This paper demonstrates that by using a nearest neighbor forwarding strategy with power control, random wireless networks can achieve the theoretical upper bound on throughput while maintaining connectivity, resolving the traditional trade-off.

Contribution

It introduces a strategy combining power control and nearest neighbor forwarding to attain optimal throughput and connectivity in random networks.

Findings

Achieves the upper bound on throughput in random networks.

Ensures asymptotic connectivity with the proposed strategy.

Optimizes throughput-delay trade-off in large-scale networks.

Abstract

The discrepancy between the upper bound on throughput in wireless networks and the throughput scaling in random networks which is also known as the connectivity-throughput trade-off is analyzed. In a random network with $\lambda$ nodes per unit area, throughput is found to scale by a factor of $\sqrt{\log{\lambda}}$ worse compared to the upper bound which is due to the uncertainty in the nodes' location. In the present model, nodes are assumed to know their geographical location and to employ power control, which we understand as an additional degree of freedom to improve network performance. The expected throughput-progress and the expected packet delay normalized to the one-hop progress are chosen as performance metrics. These metrics are investigated for a nearest neighbor forwarding strategy, which benefits from power control by reducing transmission power and, hence spatial contention. It is shown that the connectivity-throughput trade-off can be resolved if nodes employ a nearest neighbor forwarding strategy, achieving the upper bound on throughput on average also in a random network while ensuring asymptotic connectivity. In this case, the optimal throughput-delay scaling trade-off is also achieved.