The batch renewal process is the least biased choice of a process given only the measures of count and interval correlations at all lags. This article reviews the batch renewal process for modelling both LRD (long range dependent) and SRD (short range dependent) traffic flows. The exposition focuses mainly in the discrete-space discrete-time domain and in the wider context of general traffic in that domain. However, corresponding results in the continuous-time domain are also presented. Moreover, some applications of the batch renewal process in simple queues and in queueing network models are undertaken and associated analytic performance results are devised. The article concludes with open research problems and issues relating to the batch renewal process.