Protocols and minimum capacity for transmission of time-critical message in noisy channel

A time-critical message should be correctly transmitted to the receiver before a specified deadline. If the message is corrupted by transmission noise, it can be retransmitted if the resulting delay does not exceed the specified deadline. If the channel capacity is larger, the transmission time is smaller, the maximum number of retransmissions can be larger and hence the probability χ that the message can be correctly received before the deadline is larger. In this paper, we propose two protocols, called Transmit and Wait Protocol and Continuous Transmission Protocol, for handling transmission and retransmission of time-critical message in noisy channel. To ensure that χ is at least equal to a given requirement χ*, we determine the minimum capacity required by these two protocols for two noisy channel models: (1) independent error model and (2) bursty (or Markovian) error model. The Transmit and Wait Protocol is very simple but the Continuous Transmission Protocol requires smaller capacity.