[Essay draft 0.1]

The Disaster Stack

(c) 2012 by Vernor Vinge
cc-by-sa

This work is licensed under a
Creative Commons Attribution-ShareAlike 3.0 Unported License

It's a truism that technological progress increases the possibilities for disaster and makes it easier for disasters to grow into catastrophes. At the same time, there are aspects of technology that could help us deal with threats. This essay is about using hobbyists, charity, clever business practices, and government cooperation to create a "Disaster Stack" of machines, knowledge, and people to respond to disasters:

• Layer 1: The communication network.
  The two most obvious disaster threats to communications are:
  • Loss of electrical power. The power needed to run smartphones is orders of magnitude less than the needs of civilization as a whole. With a little bit of forethought and regulatory sympathy, car batteries could power such devices for some hours, and existing green sources could provide longer support. Base stations and backhaul technology have more concentrated power needs, but one interesting trend of the last few years has been use of smaller and smaller stations. Emergency backhaul may have its own power supply surprises.
  • Congestion arising from the disaster-related demand surge. Falling back to lower bitrates and datagram-oriented transport can help with surge problems. Again, there are tech trends that may make this problem more easily solvable: In principle, wireless comms can maintain high-quality peer-to-peer contact at very great station densities.
• Layer 2: The knowledge and program base that runs atop Layer 1.
  A cliche of catastrophe science fiction is the notion of a cache of reference books that explain technology from before the "fall of civilization". The cache might be as simple as an engineering manual or a survivalist's recipe list. Old-time science-fiction fans debated which 20 pounds of reference books would have the greatest payoff. Of course, nowadays we can do much better: our smartphones have enough storage to hold entire libraries. Besides storing knowledge, we can store plans and programs, customized for each of the disaster scenarios we consider.
  • Building Layer 2 would be a crowdsourced version of scenario-based planning. It would be a vast project, though in the beginning not a great deal more ambitious than Wikipedia. Even more than Wikipedia, it would be an ongoing effort: a growing hierarchy, its roots being grand categories of disaster, its leaves being discussions of particular possibilities and responses (and pointers to common response libraries).
  • One of the virtues of scenario-based disaster planning is that it allows the independent study of completely contradictory policies. (For instance, the recommended response to aircraft hi-jacking before 9/11 was very different than after 9/11.) In some cases, these contradictory policies can't be resolved before the event. Having both represented in Layer 2 would give responders insights and options (and even diagnostics) that could be applied immediately in the presence of surprises.
  • Building the Disaster Stack involves doing things long before any particular disaster occurs. Unfortunately, the number of possible disaster scenarios collides with the bounded resources of the planners (be they individuals or nation states). The structure of Layer 2 would provide insight to those with access to money -- both to inspire them to re-estimate risks and to survey the hierarchy for feasible projects that would have leverage across a range of possible disasters. So for instance, introducing a small change in smartphone sensors and protocols might have a large positive effect for many different scenarios.
  • Certification and assessment is an essential part of building the knowledge base and programs that make up Layer 2. However, considering the size and contentious nature of the possibilities, certification and assessment should not be the monopoly of any single entity. (I would prefer to have choices about whom to trust per scenario and context.)
  • Disasters will happen, but hopefully the smaller ones will continue to be more common than the larger. In principle, the Disaster Stack scheme scales from "disasters" as small as losing one's car keys all the way up to a giant meteor strike. Building Layer 2 is a neverending hobby, partly because we'll always be imagining new disasters -- but also because each disaster that really happens will give us an opportunity revise and extend Layer 2.
• Layer 3: People helping out during disasters, supported by Layer 2.
  David Brin has written much about the importance of citizen involvement in disaster training and response. He has said that while great courage and sacrifice was shown by rescue workers in the 9/11 disaster, that the most effective work was done on the spur of the moment by civilians; furthermore, the only people to recognize and defend against the attack in real time were the civilians aboard United Flight 93, aided by phone communications. Such is the inspiration for Layer 3 of the Disaster Stack:
  • At "run time" -- that is, when a disaster strikes -- the people can turn to Layer 2 services. Layer 2 will be running on top of the best available communication network (Layer 1).
  • The Layer 2 computation and data can be used to run programs using smart phone sensors to determine the nature of the disaster. The people (Layer 3) use that information to progress down the Layer 2 hierarchy to the nodes that best fit the information available. In those scenario nodes they should find advice about possible actions to take (for instance, in the case of a train wreck, the location of fire fighting equipment, medical gear, current location of medically knowledgeable persons, and so on).
  • Especially in the early stages of the disaster, different scenarios may fit the known facts. The people of Layer 3 may have to conduct further observations to distinguish the possibilities or invent new explanations for the event. (Is the train crash caused by an earthquake, or terrorist action, or mis-scheduling, or a gas leak explosion, or ...?) These results are communicated and acted upon.
  • For small disasters, the actions of Layer 3 merge seamlessly back into the everyday activities of society, including appropriate plan-time changes to all three layers. For large disasters, the Disaster Stack eventually hands off to conventional recovery institutions. For the largest disasters, the Disaster Stack might best be regarded in the spirit of the Long Now Foundation.

The dual use nature of technology is one of the most dread facts of our time. Civilization seems to be in a race between tech that could empower and protect us, and tech that would give anybody having a "bad hair day" the ability to blow up civilization. The Breakers of the world have a great advantage, since it is so much easier to break things than it is to make things. The Breakers may yet win -- we don't know just what can be invented. But I think the rest of the world, the Makers, have a chance. The Makers include the great and famous, but they also include people who simply have their own families, their own interests -- including millions of different kinds of specialized expertise. Over the last ten years, the rise of cellphones and smartphones and social media and services such as Wikipedia have convinced me that the Makers of the world have their own
advantage. There are billions of Makers, and information technology has transformed those billions into an intellectual institution that trumps all human institutions of the past. That intelligence and vigilence and good will has the potential to outmatch the destructive power of technology arising from accidents and small numbers of Breakers.

It's not surprising that many of the plan time activities for the Disaster Stack look like the research behavior of Bad Guys. This fact creates an awful temptation for otherwise reasonable governments to smother disaster planning/recovery hobbyists. Considering the fact that the Makers are our best hope against the Breakers, I hope that the temptation to suppress can be resisted and that governments come to value and collaborate with the Makers.

This Disaster Stack essay is an expansion of notes I used for a presentation at Sci Foo 12. Sci Foo turned out to be an ideal place to loft this sort of notion:

• A good many Sci Foo presentations were about projects that were within tweaking distance of being relevant to disaster planning and recovery -- and most of the presentations were about real projects and prototypes, not the handwaving of this Disaster Stack proposal.
• Between presentations, I had the opportunity to chat with other attenders about disaster issues. Stewart Brand had personal experience with the 1989 Loma Prieta Earthquake and later wrote a report about the rescuers and rescued. As with David Brin's comments on 9/11, ad hoc civilian response was very important and positive. Not surprisingly, on-duty police and professional rescuers arrived after the civilians. The police permitted the civilians to continue to participate. Volunteers plus institutional resources (fire boats and hoses) may have prevented far greater distruction. Civilians with prior experience (military, firefighting, even teaching) tended to do well, but almost all were maneuvering in perilous ignorance. Some civilians were totally clueless, more an obstacle than anything else. (I suspect I would have been in this category! And yet, the beauty of crowdsourcing is that it permits our different talents and weaknesses to be used where they can be the most constructive; I imagine that there are many people who are terrible at run time, but who could still help with the building of the Disaster Stack.)
• One participant at my Sci Foo session made the point that [however much we talk about these things] most people don't know what to do in a disaster, not even how to protect themselves. The hobbies and effort associated with building the Disaster Stack might go partway to helping with this problem, but effective behavior in most disasters comes down to certain standard field skills -- skills that are taught by the Red Cross and other organizations. Learning these skills remains as important as ever, and should be complementary to Disaster Stack planning.
• The folks attending my session raised a more ominous concern. The Disaster Stack depends on the good-heartedness and cooperation of the people involved. And yet, from The Lord of the Flies to Mad Max, our popular culture preaches that large disasters bring out the worst in us, with groups playing zero-sum and even negative-sum games. Is this so? Is it possible to rate this risk per disaster scenario? Civilization's most precious infrastructure is the framework of trust and understanding that invisibly makes all the rest possible. In losing that infrastructure, we might face a long detour back into negative sum games. On the scale of the Long Now, I think we'd come back fairly fast. Just knowing that something can be done is a powerful enabler. And positive sum games are our stairway to greatness.

(Thanks to Mike Gannis for his post-Sci Foo critique of this essay.)