18 March 2010

System of Systems

Whenever I hear someone use the phrase "system of systems" (SoS), I like to ask them to explain the difference between a "system" and a "system of systems." I've been doing this for a pretty long time now, and just this week I finally got a good (albeit paradoxical) answer.

Most attempts to distinguish between a system and a SoS fall into a recursive, almost fractal-esque jumble of un-clarity. For example, is an aircraft a system that is part of a larger SoS (which includes air traffic control, radar sites, communications links, etc), or is the aircraft itself a SoS, made up of the life support system, navigation system, weapon system (if it's a military jet)? Depends on your perspective, right? And if both terms can be applied to the same object, we might ask what the terms really mean.

It seems to me that we could simply describe any given SoS as a system. This lack of clarity in distinguishing between the two labels led me to suspect that SoS don't really exist in a meaningful way. However, after talking with a new BFF from Auburn University, now I'm thinking that SoS do exist, but discrete systems don't (thus the paradox).

Clearly the intent of the term SoS is to highlight the interrelated nature of technology systems. We use that term to remind ourselves that no system is an island. This awareness helps shape our decisions about investments, modifications, operations, etc. And along with the idea that no system is an island, I'm coming to suspect that no system is a system. It's all part of a bigger system of systems.

Or maybe my original opinion was right, and it would be simpler to just use the word "system." Thoughts?


Mark said...

Well, glad that's all cleared up. ;)

Paul said...

Dan, thanks for this blog, it's really a pleasure to read.

Whenever I read "System of Systems", in my mind I substitute "System of Heterogeneous, Kludgy, Divergent, or Legacy Systems Not Originally Intended for Interoperability." Which would set SoS apart from an aircraft, because the aircraft was, in theory, assembled from components intended for interoperability.

But SoS seems a much more concise term. It is also a very impressive title to place on a slide with a SV-1 diagram.

The Dan Ward said...

@Paul - thanks for the comment & the kind words! I like your def of SoS - definitely a helpful insight!

Glen B. Alleman said...


How about the definitions and guide from the DoD?


The Dan Ward said...

@Glen - You'd be amazed (or maybe not) by how many DoD people use the term SoS and have no idea there's a guide, let alone what the guide says.

But I have to admit I found the guide's distinctions a bit obtuse. Table 2.1 is a good start, but it's not terribly conducive for ordinary conversation. The intent behind my ongoing experiment / query is to find out if people can define the terms they use so glibly...

Glen B. Alleman said...

Sad isn't it. With a MSSM (1980), the notion of "engineering the system," is part of the language in only a few places. Some clients "get it" some don't know what we're talking about.

John Gall's "System Bible" is the book I hand out to those not getting it.