Erasure broadcast networks with receiver caching

We study the capacity of a broadcast packet-erasure network with receiver caching. The receivers in the network are divided into two groups: A group of strong receivers with small packet erasure probabilities, and a group of weak receivers with large packet erasure probabilities. The weak receivers are provided with local cache memories as compensation for their poor channels. Achievable (lower) and converse (upper) bounds for the optimal capacity-memory tradeoff are derived. The lower bounds are proved using new joint cache-channel coding schemes that significantly outperform naive separate cache-channel coding schemes. For the case of two receivers, the capacity-memory tradeoff is completely characterized for a range of useful cache memory sizes.