Fair rate allocation for broadcast channel with confidential messages

We consider a degraded block-fading broadcast channel, in which we treat each user as an “internal eavesdropper” for the information transmitted to another user. We addressed the problem of maximizing cumulative network utility, which is a function instantaneous rate of confidential data served for all users. Then, we propose a low complexity, provably optimal algorithm that jointly selects the amount of information to be transmitted to all users from the region of achievable open and secrecy rates in each block. In the described setup, there exists only one user with positive instantaneous secrecy rate, i.e., full equivocation from other users in a given block. Unlike existing scheduling approaches that serve only the user with the positive secrecy rate in each block, we propose a system that uses part of the secrecy rate for key generation, to be utilized at later times in the form of a one-time pad. To be able to serve a larger region of open-secrecy rates so that multiple users can simultaneously transmit data with full equivocation, we use superposition encoding and binning. Our system leads to a smoother service process, which increases the achieved long term utility, compared to the opportunistic scheduling approach.