Coded Parallel Transmission for Half-Duplex Distributed Computing

This work studies a general distributed coded computing system based on the MapReduce-type framework, where nodes within half-duplex network wish to compute multiple output functions. We first introduce definition of communication delay characterize time cost during date shuffle phase, and then propose novel coding strategy that enables parallel transmission among computation by delicately designing data placement, message symbols encoding, shuffling, decoding. Compared (CDC) scheme proposed Li et al., significantly reduces delay, in particular when load is relatively smaller than number K. Moreover, CDC monotonically increasing function K, while our decreases as K increases, indicating can make better use resources.