A Conceptual Model of the Effect of Diverse Methodologies on Coincident Failures in Multi-Version Software

Abstract

Eckhardt and Lee have shown that, in a precisely defined sense, the independent development of multi-version software cannot result in independence of failure behaviour. We have shown in earlier work that the use of diverse methodologies (forced diversity) for the development of the several versions may overcome this problem. Indeed, it is theoretically possible to obtain versions which exhibit better than independent behaviour. In this paper we try to formalise the notion of methodological diversity by considering the sequence of decision outcomes which comprises a methodology. We show that diversity of decisions implies likely diversity of behaviour for the different versions developed under such forced diversity. We define a measure of diversity of methodologies and show that there are simple orderings in the behavioural diversity resulting from the particular choices in design decisions. In particular it is possible to make design decisions in such a way as to optimise diversity and so minimise the chance of coincident version failure.