Trade-Offs in Reliability, Latency, and Energy for Random Network Coding-Enabled Networks

Abstract

In random network coding (RNC)-enabled networks, a sufficient number of linearly independent encoded packets, i.e., generations, is guaranteed at the cost of high latency and energy consumption at the transmitter. This letter aims to reduce the weighted sum of latency and energy while guaranteeing a targeted successful transmission probability, which is defined as the probability that the RNC-enabled network has sufficient generations when the transmitter receives a required number of acknowledgements (ACKs) feedback. An analytical framework and a deep learning-based method are developed. Results show that the latency and energy consumption are reduced by jointly optimizing the required number of ACKs and the transmit power at the transmitter.