Resolving the connectivity-throughput trade-off in random networks

摘要

The discrepancy between the upper bound on throughput scaling in wireless networks and the throughput scaling in random networks is due to the connectivity-throughput trade-off. In a random network with λ nodes, throughput scales by a factor of √logλ worse compared to the upper bound because of the uncertainty in the nodes'' location. In this work, nodes know their location and employ power control. The expected throughput-progress is analyzed for a nearest neighbor forwarding strategy, which benefits from power control by reducing spatial contention. By this, the connectivity-throughput trade-off is resolved and the upper bound is achieved.