Systems of Systems: Scaling Up the Development Process
journal contributionposted on 01.08.2006 by Watts S. Humphrey
Any type of content formally published in an academic journal, usually following a peer-review process.
Some systems have some but not all properties of systems of systems (SoS). We refer to these as SoS-like systems. This report reviews the fundamental process and project-management problems of large-scale SoS-like programs and outlines steps to address these problems. The report has eight sections. Section 1 summarizes current thinking on the nature of future complex systems, and Section 2 discusses the systems-design problems of the future, particularly the partitioning of massive systems into system-of-systems structures. Section 3 points out how large-scale systems development efforts have typically failed because of project-management and not technical problems, and that the solutions to these problems are known and highly effective, but not widely practiced. It explains why, if the project-management problems of the past are not promptly and effectively addressed, large-scale systems development programs will likely be unmanageable. Section 4 discusses the requirements for a scalable process, and Section 5 both reviews and explains the quality-management principles upon which any scalable process must rest. Section 6 reviews the nature of the project-management problems currently faced by large-scale software-intensive system development efforts and explains why attempts to scale up current methods to very large-scale systems work will almost certainly fail. Section 7 describes process strategies for supporting development of a network-like system of systems and it outlines the process and project-management topics needing further research and development. Finally, Section 8 reviews the process considerations for supporting the very large-scale integrated development programs of the future. The report concludes that, unless steps like those outlined in this report are taken in conjunction with continuing technical research and development, the large-scale systems development efforts of the future will almost certainly fail, and often catastrophically.