An interference-aware perspective on decoding power

Traditionally, coding has been seen as a way of saving transmit power: capacity-approaching codes require minimal transmitted energy-per-bit given the bandwidth available. But because transmit power is often smaller than decoding power at short distances, many recent wireless system designs continue to use uncoded transmission! We first observe that in wireless systems that both generate and face interference, coding serves another purpose (assuming interference is treated as noise): it allows a system to support a higher density of transmitter-receiver pairs. Bringing decoding power into the picture, we propose an approach to investigate which code/decoder to use and whether to use any coding at all. It turns out that the code´s gap to capacity determines how high the maximum supportable link density can be when power is plentiful, whereas the code´s decoding complexity governs what link densities can be supported at low power.