Throughput, Delay, and Complexity Tradeoffs in Interference Channels

We investigate the duality of the stable throughput and the capacity regions in the context of networks with distributed transmitters. In particular, we focus on a network with two transmitter-receiver pairs, intermittent connectivity, and delayed feedback. For this problem, we show the equivalence of stable throughput and capacity regions by devising a queue-based dynamic communication protocol that achieves the information-theoretic outer-bounds. This new strategy includes several new ingredients and coding strategies compared to prior results. We assume packets arrive through stochastic processes and explain how our procedure determines packet movement rules between queues in different cases. We also investigate the average lifetime packets spend in the queues under this capacity achieving protocol and observe that the lifetime scales the square root function of the communication window.