Allocating data for broadcasting over wireless channels subject to transmission errors

Broadcasting is an efficient and scalable way of transmitting data over wireless channels to an unlimited number of clients. In this paper the problem of allocating data to multiple channels is studied, assuming flat data scheduling per channel and the presence of unrecoverable channel transmission errors. The objective is that of minimizing the average expected delay experienced by the clients. Two different channel error models are considered: the Bernoulli model and the simplified Gilbert-Elliot one. In the former model, each packet transmission has the same probability to fail and each transmission error is independent from the others. In the latter one, bursts of erroneous or error-free packet transmissions due to wireless fading channels are modeled. For both channel error models and unit length data, an optimal solution can be found in polynomial time either when all the channels have the same error probabilities or when there are exactly two channels with different error probabilities. In the remaining cases, including non-unit length data and error probabilities which differ from channel to channel, sub-optimal solutions can be found for both error models. For these cases, extensive simulations, performed on benchmarks whose item popularities follow Zipf distributions, show that good sub-optimal solutions are found.