One might expect the reliability of a distributed system to correspond directly to the reliability of its constituents, but this is not always the case. The mechanisms used to structure a distributed system and to implement cooperation between components play a vital role in determining the reliability of the system. Many contemporary distributed operating systems have placed emphasis on com- munication performance, overlooking the need for tools to integrate com- ponents into a reliable whole. The communication primitives supported give generally reliable behavior, but exhibit problematic semantics when transient failures or system config- uration changes occur. The resulting building blocks are, therefore, unsuit- able for facilitating the construction of systems where reliability is important.

This article reviews 10 years of research on ISIS, a system that pro- vides tools to support the construc- tion of reliable distributed software. The thesis underlying ISIS is that development of reliable distributed software can be simplified using pro- cess groups and group programming tools. This article describes the ap- proach taken, surveys the system, and discusses experiences with real applications.