Evaluating System Reliability from the Customer Perspective to Improve Availability Predictions

In system availability engineering, industry focus has typically been on preventing underlying hardware and software defects and predicting system reliability during the development cycle. Most reliability practitioners have paid less attention to system level reliabillity assessment from field experiences. In this paper, we propose shifting the focus of analysis and prediction by using methods for achieving data-driven, customer focused reliability and availability assessment. Such methods are not often discussed in the literature. In particular, we show how to successfully use the field data of existing products or product releases to baseline the system availability and estimate key reliability modeling parameters. We also present how to calibrate the availability assessed from the field into an availability prediction of a new product or a new release. We illustrate our methods with a case study, and offer suggestions for further work that will extend the practicality and applicability of our methods.