Thursday, September 21, 2006

Systems of Systems

At the ACT joint systems interest groups (SQA, SESA, IEEE-CS and SCSC) meeting this week Shaun Wilson, from Aerospace Concepts Pty Ltd www.concepts.aero talked about "Challenges of Socio-technical Systems-of-Systems Engineering":

The World is full of systems, natural and man-made. We all know this; indeed, this conference and the Systems Engineering Society of Australia exist to further the practice of engineering complex systems because history has demonstrated that design and management approaches that explicitly recognise the nature of systems are useful in undertaking man's more challenging technical projects.

Recently, and especially in the Defence sector, the term 'systems of systems' has become commonly heard, perhaps to the point of being over-used. What is the difference between 'SoS' and systems in general. And what about the socio-technical aspects of SoS? To some, there is no fundamental difference, it's all just a matter of scale ... but is this really the case?
I had difficulty understanding what "Systems of Systems" were. A quick search of the Wikipedia gave:
"System-of-Systems is a relatively new term that is being applied primarily to government projects for addressing large scale inter-disciplinary problems with multiple, heterogeneous, distributed systems that are embedded in networks at multiple levels and multiple domains. ..."
Almost all of the audience had a Defence Department background, including myself. Shaun drew his examples of systems of systems from defence and aerospace. However, as an IT person, the idea of having to link to together existing systems just seemed to the everyday routine, not a new concept. With the advent of the Internet and the web, having to get existing systems (and the people who build them) to work together is not unusual. It may be new to systems engineers, but is not to IT ones.

There seems a fundamental flaw in the System-of-Systems concept. It tries to fix problems with large defence project from an engineering point of view. But many of the problems with Defence projects are not due to a deficiency in engineering or systems design, they are from political pressures preventing effective management. The engineers know what needs to be done, but bureaucratic impediments prevent them doing it. To hide the management problem is in the quasi-engineering jargon of "socio-technical aspects" is the wrong approach.

An example I gave in a question was the color of the engine nacelles on the Boeing 787 Dreamliner airliner. Boeing has proposed to airlines that the nacelles be painted a standard shade of grey. This allows the engines to have a very smooth paint job, reducing drag and saving fuel. But without the financial incentive, there would be no reason for the airlines to agree to do this. In the defence environment there would not be this financial incentive and there would have to be rules and committees invented to try to have the color standardized. The solution is to introduce the financial incentive and let the projects respond to that incentive.

An example Shaun used was the major upgrade problems for the US Air Force to keep B-52 bombers flying for 90 years. His solution was to use systems-of-systems techniques. This seemed to me to not be an engineering problem at all, but a simple case of bad management. The solution is to get rid of the old planes and buy some new ones.

The cost of developing a large military aircraft to replace the B-52 would be prohibitive, so the alternative would be to militarize commercial airliners. That might sound hard to do, but the US Navy is converting the 737 airliner into the armed P-8A Multimission Maritime Aircraft (MMA).

The B-52 is comparable in size and payload to a 787 airliner. It would take a little mechanical work and a lot of software to turn a 787 into a bomber. But at least they would not have to repaint the engines, as the standard grey would do fine. ;-)

ps: The Australian military is unlikely to buy any 787 bombers. But they are buying some C-17 transport aircraft which could be used in the role. This might just need a software upgrade. The C-17 can launch a missile 20 m long of 27,000kg.

No comments: