Coded Computation Against Distributed Straggling Decoders for Gaussian Channels in C-RAN
share
The uplink via additive white Gaussian noise channels of a cloud radio access network (C-RAN) architecture is studied, where decoding at the cloud takes place over network function virtualization on commercial off-the-shelf servers. To mitigate the impact of straggling decoders, the cloud re-encodes the received frames via a linear code before distributing them to the decoding processors. In this paper, the computation rate is derived such that all the servers can recover the desired linear combinations of the messages with vanishing average error probability. It follows a computation of an upper bound on the decoding error probability of each server. Finally, two analytical upper bounds on the frame error rate (FER) as a function of the decoding latency are developed, in order to quantify the tradeoff between FER and decoding latency at the cloud.
READ FULL TEXT