Reducing system software project risk through choice of project architecture

Abstract

The choice of project architecture - the structure of and interrelationships between product, processes, and organization - alters the project's risk profile. While most analyses take project scope as an input, I propose the examination of multiple project decompositions take place as part of project planning and project monitoring. The sub­projects created by each decomposition will have unique risk profiles, suggesting different process and organizational adaptations that lower overall project risk. By selecting project decompositions that partition risk and then adapting the structure of each sub-project to mitigate its particular risks, the probability of risk occurrence is reduced and the severity of consequences may be reduced. Case studies of four IBM mainframe system software projects illustrate lessons regarding project architecture, some general and some project- or process-specific. These projects employ both waterfall and iterative process models, managed using varying degrees of functional, lightweight, and heavyweight organizations.