Servicing Losses in Large Data Collection Systems

This paper presents an analysis of large information collection systems. The method employed consists of evaluating the probability of the system losing information due to overloading the receiving system with signals to be processed. Performance characteristics are derived that relate the probability of lost or incompletely processed signals versus the number of servicing receivers and mean arrival rate of the signals. Also considered is a special case of data redundancy in which the signal is allowed to repeat itself a fixed number of times. For this case, the probability of an incompletely serviced signal is related to the number of repeated data intervals and mean arrival rate. Most of the results are obtained analytically and are verified by the results of a Monte Carlo simulation using a digital computer.