A Measure of Guaranteed Availability and its Numerical Evaluation

Abstract

Average system availability measures such as average availability over a time interval and steady-state availability are common system evaluation measures. Recently, many commercial computer systems have offered guaranteed availability levels over finite observation periods. Average availability measures are not adequate when there is a penalty associated with not meeting this guaranteed level of availability over the specified time interval. In such a case, the evaluation of the probability of not meeting this availability requirement becomes necessary. In order to assess the goodness of the offered guarantee, we propose a success (risk) measure of guaranteed availability. Using a generic system model, we describe the measure and study the effects of the guaranteed level and the observation period on it. Furthermore, we introduce a numerical approach for continuous time Markov chain models which allows component level modeling, Coxian failure and repair distributions, time-dependent failure and repair rates, and deferred repair and nondeferred repair strategies to be handled. An example of a fault-tolerant database computer system is considered and the measure of guaranteed availability is evaluated for various guaranteed levels and observation periods. © 1988 IEEE