A New Approach to System-Availability Analysis

This paper describes a new approach to system-availability analysis. The purpose of this work is to gain a more comprehensive understanding of the contributors to system downtime through detailed modeling, with the eventual goal of reducing total cost of ownership of the system by improving availability. We developed an integrated system-availability model that accounts for all major supplier-attributable root causes of failure, hardware, software, and procedures, as well as their interactions. In previous approaches to predicting system availability, hardware and software have typically been modeled independently, partially combined, or combined at a higher level; procedures have frequently not been modeled. We applied the new approach to an optical telecommunications transmission system. Downtime predictions from the model are in reasonable agreement with estimates from field data. This model extends previous approaches by capturing interactions among root causes and offering a holistic view of system availability from the customer's perspective. The model can be used to enhance system availability by identifying and reducing primary contributors to downtime. Analysis can be performed during the early development phase to assess and enhance reliability architecture. The approach can be generalized to other types of systems. 1. Introduction Converged communication systems are becoming increasingly feature rich. Interactions among components are growing in complexity, and involve software (SW) (e.g., application and control), hardware (HW); and procedures (e.g., user-interface) layers.