A simple estimation method of network reliability with failure scale

The service unavailability has widely been used as ameasure of the reliability of communications networks. However, the previously used service unavailability is a measure of either a specific or an arbitrary user, so it cannot take account of the scale of the users affected by the failure. In this paper, we introduce a failure scale parameter as a measure of reliability called the unavailability, which enables the actual value of the reliability of a communications network to be easily estimated. Evaluation examples obtained by applying this method to the failure analysis of an actual communications network are presented. This method is shown to be effective in terms of being compatible with the existing unavailability and reducing computational effort. Furthermore, its estimation accuracy is quantitatively verified and its range of applicability is clarified. (C) 2006 Wiley Periodicals, Inc.