Wednesday, October 27, 2010

Centralized vs distributed organisms

This article from Techdirt contemplates the nature of centralized vs distributed organisms, in the context of Wikileak's recent exposures of the U.S. Government's behavior in Iraq. Regardless of one's opinion of this particular situation, it's worth reading because it provides food for thought about what our future may look like.

This is a topic that, through my life, I have thought about in different, specific, contexts - not least of which is how the pendulum has swung over the years with regard to the pros and cons of server- and client-side computing. Central came first with mainframes and dumb terminals. Then minicomputers came on the scene, reducing the cost - which led to "microcomputers" (who among us remembers that term?), aka PCs. And now the pendulum is swinging back, with all our data (and intelligence) in the "cloud."

Perhaps a good illustration of the way this works is the Microsoft/Google analogy. Microsoft built its empire on the notion of putting the smarts into PCs by having you buy their software, and then locking you into occasional "upgrades" - the distributed model. Google came along later, and realized that a centralized model made more sense. So now, Microsoft is in the sad position of having to release updates of its software (e.g. Outlook) on a periodic basis. Otoh, Google, by virtue of the fact that its software (e.g. Gmail) is in the cloud, is in the much happier position of being able to control updates centrally, as and when they want or need to. If you're a user of Google's products, you will notice how they "push" subtle changes to you on a regular basis. Microsoft wishes they had the ability to do this. But that's not really what this post is about, and I happily recognize that I've elided over many points that make the computing example much more complicated.

Back to centralized vs distributed organisms... Beyond computing, where else does this occur?

Well, there's the Cold War, where the enemies were easy to define: sovereign nations. With the Cold War behind us, we now have global terrorism to worry about. Now, the enemy is much harder to define, because instead of countries with clear boundaries, we find ourselves battling ideologies. Small cells of trouble can pop up anywhere, and vanish just as quickly. So: Cold War - centralized. Terrorism - distributed. Is it any wonder the coalition forces going up against the likes of Al Qa'ida are having so much difficulty? We still model our defense forces on the old, centralized model, yet we're fighting a distributed "war." It needs an entirely new approach, one that will not come easily or quickly.

Another example is that of governments. Why is China gaining ground on the U.S. in oh-so-many aspects, not least of which are their respective economies? China uses an autonomous, centralized model. The U.S. has a federal system, which devolves much of the power to individual states - a distributed model. When China wants to achieve something, they simply do it, declaring it as a fait accompli. In the U.S., however, it's a much more complex process - eternal wrangling over, for example, whether to spend or save. The consequence, of course, is that very little gets accomplished here with any alacrity.

So there you have three examples: computing, war, government. I'll leave it to you, dear reader, to decide which is preferable in each case: centralized or distributed organism? And to what other situations can we apply this? Comments, as always, welcome!

No comments: