Optimum packet size for data networks with random channel errors

The formula for the optimum packet size required for maximum utilisation of network lines can be derived in several different ways, under a usual assumption of error-free transmission lines/channels. An analysis based on the concept of minimisation of network overheads is used to derive the optimum packet size for lines having random-error characteristics. The messages are assumed to have a Poisson distribution, and the network uses a stop-and-wait strategy for error detection and correction.