How Economics Shape Reliability: Lessons and Opportunities from Windows Development

Abstract

How can we deliver highly reliable software profitably, for low cost and at large scale? The value of answering this question is enormous: providing software quality costs organizations, let alone consumers, roughly half a trillion dollars per year worldwide. Large scale data suggests that discovering and correcting defects in the traditional ways, while necessary, is not a sufficient answer. This talk will review some of the salient efforts and results of the Windows team in pursuit of high reliability. We will identify some long-term challenges and describe the built-in tools and capabilities that enable the research and industrial communities to study and address reliability issues in the Windows ecosystem.