Two rapid test methods for assessing transmission time reliability of a multiple-channel network with Poisson arrival and service rates

Abstract

To shorten the duration of the network transmission time reliability test, two rapid methods are proposed for a multiple-channel network with Poisson arrival and service rates. For a network whose service rate is given and can be increased in practice, a rapid test can be conducted on a similar network with higher arrival and service rates. The test conditions are derived using the similarity theory, and the data collected in the test can be used to estimate the reliability of the original network. For a network whose service rate is unknown, a rapid test analogous to accelerated test can be applied, which only increases the arrival rate. The analytical expression of transmission time reliability is used to derive the accelerated model, and the data collected in the test can be used to extrapolate the network reliability under normal arrival rate. Our numerical study shows that the two proposed methods are efficient and the reliability estimates are quite accurate compared to the empirical estimate obtained from verification tests. In the rapid tests, the data collected in a given period of time increases along with the arrival rate, and the test duration can be shortened without affecting the sample size of packets.