I'm trying to write a post on resilience, complexity, and (perhaps) self-organizing criticality and I'm looking for an example of a complex system that is or can be objectively made more resilient by decentralization. I've done a google search and haven't found anything I like.
Anybody have any ideas?
33 comments:
Cloud computing.
That's better than anything I've thought of so far.
I was hoping for something a little more familiar to a wider audience - cloud computing is kinda geeky. But it will at least be one of the examples.
Gasoline distribution.
Let's see if understand that.
So if gasoline was distributed from a single (or small number) or points (and perhaps only refined at those points), then disruptions in terms of storms, strikes, terrorist attacks, whatever, would make supply more unavailable than if refineries and distribution were all over the place?
I think that's right and I like it. Is that what you meant?
Yes. You could apply in a general way to most consumer goods (food, for instance). You could also look at decentralization in terms of organizations and not just geography (e.g., having multiple companies selling gasoline makes it more likely some will be available).
I reread what I was trying to write now that aog has given me some suggestions.
As more responsibility and the associated authority and experienced is pushed down from a centralized government to communities and individuals for their own recovery from whatever catastrophes appear, I think that they're more likely to be effective at dealing compared to when a centralized government takes care of everything (for example if FEMA actually worked).
The first metaphor for this started to be the pile of sand where dropping a single incremental grain can cause a substantial part of the sand pile to collapse (Bak–Tang–Wiesenfeld sandpile). That would be the fragile, non-resilient part of the metaphor, but it doesn't quite work and I'm not sure what the resilient part would (maybe wet sand?).
So I'm looking for a metaphor that has to do with individual responsibility, experience, and outlook. Grains of sand wouldn't be bad if I could pull it off.
links in a chain
Don't forget to mention subsidiarity.
Perhaps a highway / road analogy? Suppose we had only highways, but really wide ones - would that be more resilient?
Or maybe a football teams vs. Madden 2012 - which will react better to the unexpected, one guy running the whole team with a controller, or 11 independent intelligences operating on the field? Or a variant, the cooperative group that divides tasks vs. the group run a controlling personality who must decide on everything?
A convoy vs. a supertanker?
Like the Borg? Roddenberry sure had a handle on socialism!
Lefties can't distinguish between fiats from a ruling elite and independent free people working together to accomplish a task.
Lefties also hate highways and people with their own vehicles more than anything else -- makes them, like cats, very hard to herd.
How about the Internet? The more nodes, the more datacenters, the more cables, the more resilient. In fact, there is no centralization for the net, really, save for the TCP/IP set of rules (da law!).
jeffs -- ya think that's why it works so well and why the elites want to regulate it to make fairer?
Hmmm. I can think of a counterexample. A few years ago, I reviewed a book that explored why local public health measures failed but national measures worked when the problem was yellow fever:
Yellow Fever and the South (Paperback)
"Yellow Fever and the South" is not about the suffering and death from the repeated visits of this terrifying disease, although it is about the panic. The suffering and death can be found in other histories.
Humphreys, a medical doctor, wrote this dissertation for a history degree, and it traces the evolution of public health organization in the South from the Civil War to and a little beyond the last yellow fever epidemic in the United States in 1905.
It is full of surprises.
Yellow fever was not the greatest killer in the South. TB, typhoid, malaria, and the debilitations of pellagra, hunger, hookworm etc. killed more people. But those deaths were background and people took them as they came. Yellow fever visited sometimes every summer, sometimes not once in a decade.
When it did, though, it spread panic. Business was almost shut down. It is Humphreys' contention, no doubt correct, that it was the interruption of business, not suffering and death, that inspired or forced local and state governments to found departments of public health.
That was a sword that cut two ways. Since the true vector for the fever was unknown until 1901, cities would invoke quarantine against other cities, which may not have had the disease, for commercial advantage. In the countryside, "shotgun quarantines" of panicked citizens overrode the attempts -- when they were made -- to coordinate the official measures.
Because the southerners could not trust each other, by the 1890s they were ready to turn public health over to the national government. This from men who had, in many cases, carried arms for state's rights in the `60s!
This makes for pretty dry reading, but the repayment for the effort is understanding how people, as individuals, towns, provinces and nations, react to terrifying epidemics, whether they think they have prophylactics or not. In the 21st century, this lesson has obvious import.
Unfortunately, if the time comes, the people who will be loudest and most insistent will be those who are least worth listening to. We've already had an epidemic of that with Tamiflu.
Serious people will want to know what's in this book.
Hmmm. I can think of a counterexample. A few years ago, I reviewed a book that explored why local public health measures failed but national measures worked when the problem was yellow fever:
Yellow Fever and the South (Paperback)
"Yellow Fever and the South" is not about the suffering and death from the repeated visits of this terrifying disease, although it is about the panic. The suffering and death can be found in other histories.
Humphreys, a medical doctor, wrote this dissertation for a history degree, and it traces the evolution of public health organization in the South from the Civil War to and a little beyond the last yellow fever epidemic in the United States in 1905.
It is full of surprises.
Yellow fever was not the greatest killer in the South. TB, typhoid, malaria, and the debilitations of pellagra, hunger, hookworm etc. killed more people. But those deaths were background and people took them as they came. Yellow fever visited sometimes every summer, sometimes not once in a decade.
When it did, though, it spread panic. Business was almost shut down. It is Humphreys' contention, no doubt correct, that it was the interruption of business, not suffering and death, that inspired or forced local and state governments to found departments of public health.
That was a sword that cut two ways. Since the true vector for the fever was unknown until 1901, cities would invoke quarantine against other cities, which may not have had the disease, for commercial advantage. In the countryside, "shotgun quarantines" of panicked citizens overrode the attempts -- when they were made -- to coordinate the official measures.
Because the southerners could not trust each other, by the 1890s they were ready to turn public health over to the national government. This from men who had, in many cases, carried arms for state's rights in the `60s!
This makes for pretty dry reading, but the repayment for the effort is understanding how people, as individuals, towns, provinces and nations, react to terrifying epidemics, whether they think they have prophylactics or not. In the 21st century, this lesson has obvious import.
Unfortunately, if the time comes, the people who will be loudest and most insistent will be those who are least worth listening to. We've already had an epidemic of that with Tamiflu.
Serious people will want to know what's in this book.
Interesting. So what happened after the southerners were willing to turn public health over to the national government? Did the panics go away? If so, why?
This should be good.
erp,
I often disagree with Harry, I'm sometimes unable to comprehend what he says, but I always find his answers interesting.
"It is Humphreys' contention, no doubt correct, that it was the interruption of business, not suffering and death, that inspired or forced local and state governments to found departments of public health."
If by interesting, you mean predictable, I agree. The very economic system that brought peace and prosperity to the most people, is in Harry's lexicon, the one to be most despised.
As I've said before, logic isn't Harry's first language.
One of these days I am going to write up my own real-life experience, but until then ...
The key to success is getting people to reach some reasonable consensus on what needs doing, then give them the power and responsibility to do it.
Which FEMA does not.
The key to getting things done is for one person to see a need and then take the lead and do the job with or without a consensus
Skipper who exactly gives them the power and responsibility?
Things worked great here because we had a simple blueprint and left the details to individuals who then went their own ways with only the guidelines of the law to keep us from anarchy.
All government agencies exist to continue their existence, not to do any real work or serve the public. FEMA is a perfect example.
I'm looking for an example of a complex system that is or can be objectively made more resilient by decentralization.
International emergency relief? Urban development and municipal planning? Vice regulation and control? Ordinary law enforcement?
Harry, no offence, but your example works because you are talking about a universal threat that presents itself universally and indiscriminately and can only be solved universally. Local flood control wasn't much help in Noah's time. It's like arguing that if the Martians invaded Earth, national armies in alliance under a central command would be better than local militias, so the UN should therefore take over the world's armies.
If the public health threat were local flooding or fallout from local industrial poisoning, would you not perhaps conclude differently? Do you consider FEMA your ideal theoretical protection against Hawaiian volcanos?
"It's like arguing that if the Martians invaded Earth, national armies in alliance under a central command would be better than local militias"
I've read a number of SF stories which take as a key plot point exactly the opposite - the unification of Earth's military makes it possible for the aliens to use subversion to destroy our ability to resist, making the central command much worse than national militias.
I have to admit it wasn't overwhelmingly clear to me from viewing NATO's operation and the like that global control of all military forces would be more effective against an alien invasion than independently controlled forces.
Nonetheless, I like Peter's point in that EVEN IF it would be better in the case of alien invasion, it wouldn't be a good idea in the general case.
i read those stories too as a kid and they helped shape my thinking that power should stay local. if we survive this onslaught of socialsm, which i doubt, it'll be because enough yahoos* have guns and know how to survive in the wild.
*non-intellectuals
erp,
I suspect you'd really like the blog http://accordingtohoyt.com/ . Like you, she's also quite, quite concerned by the alleged "onslaught of socialism" and I've been skewered for expressing my opinion that perhaps the concern that the onslaught is going to cause imminent collapse is a bit overwrought.
(She was one of Instapundit's co-bloggers in the weeks before the election).
I've read a number of SF stories which take as a key plot point exactly the opposite
SH, we have to talk. The key word you and erp are skimming over here is "fiction". :-)
A leftist Dutch journalist wrote a book about UN peacekeeping in the nineties. It was full of anti-American, world-government shibboleths, but she was honest enough to report the reality before her eyes. She was in Haiti during a near-civil war when a UN peacekeeping force of Euros and Canadians went it. They were the last word in politically correct, multicultural deference and respect, but they couldn't do a bloody thing to protect the frightened Haitians, provide services, etc. It was all endless coordination and administration. Finally, much to their distaste, they were replaced by U.S. Special Forces who made no bones about the fact they thought they were in a hellhole and that everybody was hopeless. But they had a job to do and a platoon leader gathered a terrified village and told them that if they stuck close to him and did exacly what he said, they'd be fine. So they stuck close to him and did exactly what he said and they were fine. They may have been hopeless, but they weren't stupid.
Nonetheless, Bret, you may want to be careful about using that tale to defend sending the Green Mountain Boys up against the Martians.
perhaps it's generational. my world was far different and not only technologically. the main difference as i see it, is that the dole has been turned into entitlements.
we've traded self-respect for trinkets and that has turned us into willing slaves, serfs, cogs, call it whatever you will. we now accept the state as the master, not the servant it should be.
i'm 78 and have no desire to live long enough to see what i'm pretty sure is coming.
btw - i read few blogs and except for you guys have no interest in the opinions of others.
pretty much heard and seen it all.y
Peter;
Well stated, but until we are actually invaded twice by aliens (once with a unified military, once without) it's all fiction.
Mainly though I was just poking at you - Bret's point is the serious take on the subject.
In the event, yellow fever was eradicated immediately after the federal government took over.
I am not against decentralized, resilient systems and am puzzled why it was hard to think of one.
Electrical systems are designed that way. Hawaiian Telephone Co., which faced lots of different ways to have its interisland network disrupted (fire, storm etc.) used a network of semi-independent relays of different type (digital radio, line-of-sight etc.) that was highly redundant.
When an earthquake knocked cell towers off the air, Hawaiian Telcom didn't miss a beat.
The problem -- or a problem -- with decentralized networks is maintenance.
Decentralized maintenance tends to be more expensive and to get skipped more often than central maintenance.
That's been my objection to rooftop solar. Installers tell me that after 10 years, only 90% of Hawaii installations are still working. Maui Electric Co., on the other hand, has generators in service that are older than I am.
"I am not against decentralized, resilient systems"
I have yet to see any policy proposal from you that doesn't discard decentralized resilient systems for centralization. Such as, say, health care. So you're not against them in general, just in every specific instance. Got it.
[erp:] Skipper who exactly gives them the power and responsibility?
To be extremely brief, I inherited responsibility for hurricane evacuation planning when I was stationed in FL. The plan I got handed was very centralized, despite the fact there were at least 15 different organizations involved.
At staff meeting shortly after I got this thing, I voiced my opinion that this long-standing plan had significant problems with decision making and execution because the people who were responsible had no authority.
Fortunately for me, a month later a Cat-1 hurricane came through. Since our facilities were rated to withstand a Cat-1, we only needed to do a minimal evacuation (five airplanes couldn't fit in the available hangar space). Despite that, it was akin to a pack of monkeys fighting over a football.
At the following staff meeting, the Commodore decreed that, in matters of hurricane evacuations, I spoke for him.
Whereupon I told each of the organization heads -- some of which had conflicting requirements -- that I needed a plan from them on what they needed to get their job done. My job was to run a sanity check on the results, and resolve conflicts.
Luckily for me, a second, much larger hurricane came through a couple months later. Once the decision was made to execute the plan (which now gave the weather folks the hammer), each of the various agencies executed their own plans, for which they were responsible.
It went perfectly.
Moral of the story: there valid roles for central authority -- coordination and arms-length sanity checking being chief among them.
Unfortunately, FEMA (and a whole bunch of other government agencies just don't get that). The moment FEMA takes over execution, then widespread passivity and dependency becomes the rule.
FEMA could be useful as the Federal Emergency Planning Agency, which would review state plans for adequacy, and coordinate recovery resources regionally, and distribute after-disaster lessons learned.
Such an agency would be a heck of a lot smaller than FEMA.
(I hope that wasn't too OT. Looking at Bret's post just now, it is. Sorry, but I'm not about to bif that many bits.)
---
Isn't the human brain (or any brain) a decentralized, resilient system?
If not, animals and plants probably are. Plants in particular don't seem to have any centrality at all.
skipper, not surprisingly, you just made my point for me. you saw the need and took the lead ... and solved the problem w/o creating a new federal program.
Thanks to all for the input. I've posted (the first version) of the result. I'll probably update it based on comments.
Happy Thanksgiving!
Post a Comment