Delay-Constrained Maximum Information Output Rate in Fading Channels

Delay and information transmission rate are two fundamental quantities in communication systems. The information-theoretic capacity provides the ultimate limit of the transmission rate for reliable communications, yet with the effect of delay either completely ignored or partially considered. How to incorporate delay constraints into capacity characterization is one of the long-standing fundamental problems in wireless communications, which has gained a renewed sense of urgency with the increasingly demanding requirements for high data rate and low latency. The challenge lies in establishing an analytical framework that can integrate the accurate characterization of delay with the information-theoretic capacity. In this paper, by modeling the information flow as a queue of codewords, we demonstrate that the service process is the key that associates both the information output rate and the access/queueing delay performance. For a given mean-access-delay constraint, the maximum information output rate is characterized in various conditions in fading channels. The analysis shows that the mean-access-delay constraint has a crucial impact on the maximum information output rate when it is below a certain threshold. To achieve the maximum information output rate, the information encoding rate needs to be carefully tuned according to the delay constraint, which diminishes as the constraint becomes more strict. Discussions on how to incorporate feedback overhead and other delay constraints including the mean queueing delay and access delay deadline into the rate analysis are also presented.