부산대학교 도서관

A model of input data buffering in measuring systems is developed. The authors obtained the formulae to determine the following time intervals: between two successive transactions to signal source, two successive transactions to signal receiver, between transactions on data receiving and generation, between transactions on data generation and receiving. It is shown that an adequate model of the buffered systems functioning is a complex Markov process, each state of which is a 2-parallel Markov process describing the “competition” between transactions filling the buffer and the transactions emptying the buffer. The complex process was transformed into the ordinary Markov process, the states of which simulate the number of currently filled buffer cells. The ordinary process analysis permits to obtain a formula interlinking the failure probability, the buffer memory volume and the ratio between the mathematical expectations of the time necessary for buffer filling and emptying.