Enumerating Software Failures to Build Dependable Distributed Applications

Abstract

We are proposing a systematic approach to building reliable distributed applications. The main objective of this approach is to consider reliability from application inception to completion, adding reliability patterns along the lifecycle and in all architectural layers. We start by enumerating the possible failures of the application, considering every activity in the use cases of the application. The identified failures are then handled by applying reliability policies. We evaluate the benefits of this approach and compare our approach to others.