We study a one-hop broadcast channel with two receivers. The receivers have side information obtained by overhearing wireless channels. The relay takes control decisions by coding transmissions based on its knowledge of side information in the receivers. We consider two control mechanisms. In the ACK system, the relay has definite knowledge of side information announced via overhearing reports. In the NACK system, the relay has statistical knowledge of side information and receives feedback after every decoding failure. Our contribution is as follows. We provide the minimal evacuation times for the two systems and obtain analytical expressions of the throughput region for the ACK and the code-constrained region for the NACK system. When the transmission rates are the same (r1= r2) or when the receiver with the highest transmission rate has perfect side information (pf=1), we show that the two regions are equal. We then provide simple joint xor coding and scheduling policies that achieve those regions and, thus, are throughput optimal. Subsequently, we evaluate the report overhead performance for both mechanisms and reflect on the involved tradeoff with throughput. Ultimately, we demonstrate by simulations that the proposed throughput optimal policies can be appropriately enhanced to have good delay properties, particularly for protocols that utilize sequenced packet delivery.

We study an 1-hop broadcast channel with two receivers. Due to overhearing channels, the receivers have side information which can be leveraged by interflow network coding techniques to provide throughput increase. In this setup, we consider two different control mechanisms, the deterministic system, where the contents of the receivers' buffers are announced to the coding node via overhearing reports and the stochastic system, where the coding node makes stochastic control decisions based on statistics and the performance is improved via NACK messages. We study the minimal evacuation times for the two systems and obtain analytical expressions of the throughput region for the deterministic and the code-constrained region for the stochastic. We show that maximum performance is achieved by simple XOR policies. For equal transmission rates r1= r2, the two regions are equal. If r1≠ r2, we showcase the tradeoff between throughput and overhead.