- From: C H <craighubleyca@yahoo.com>
- Date: Thu, 21 Jun 2007 02:35:40 -0700 (PDT)
- To: public-disaster-management-ont@w3.org
- Cc: William Waites <ww@groovy.net>
-------- SUMMARY: Regarding the "Emergency Telecommunications cluster", it seems to me that this has to be thought through a lot more carefully than it ever has before, since the ontology is meant to enable a very different type of emergency response than has been practiced previously. Newt Gingrich suggests that a modern disaster response system must resemble "a cross between E-Bay, Craig’s List, and an online dating service. Those who need help will be signaling to those who have help to give and they will be mutually interacting with minimum bureaucratic interference." He is probably right: The requests and offers can be sent by shortwave or carrier pigeon or motorcycle courier. The supplies can go by local bus or donkey train. The transactions can involve private donors, including donors of logistics and transport and storage services, and volunteers. Regardless of how they are reached or how they commit, their coordination is going to require the Internet anyway. Accordingly, the faster the call for help reaches the Internet and the later the offer or authorization to deliver certain supplies leaves the Internet, the better. Net-centric coordination may be the only way to deal with disasters that make workplaces inaccessible but leave communications in general intact (such as a pandemic or most conflicts). Relying on non-emergency-specific services and means of communication wherever possible (for logistics, for finding volunteers) may be the most practical way forward if only because it focuses the ER experts on the actual emergency, not on supply chains, not on calls for help, not on telecom or logistics. While some disaster-specific protocols may be required until communications is re-established and reliable, and may help ease the load on cell phone and satellite networks, the long term trend is towards IPv6 used on all devices from power-saving home automation systems (IEEE 802.15.4) to terabit switches. RFID and QR code and the GPS capabilities of cell phones should all be thought of as extending the reach of the IPv6 network. Because all these devices all need DC power, either from (5 volt) USB or (12 volt) vehicle-style batteries or (much less desirable) other types of batteries, the ontology must probably deal with DC power supply/load as a capability of a facility or vehicle or device or item. This could be done by importing some IEEE terms. Resilience is radically improved by relying on these methods, which use much less power and require much less travel than any method based on people moving around, such as William's need to travel to the Gulf Coast because information wasn't available from home. AC power grids probably should be considered a luxury and in any case are often the first things to fail - in many countries they're so unreliable they might as well not exist. When they work, they probably should only be used to charge batteries or stationary fuel cells, freeze things, boil water and call for more DC power sources like solar panels to be shipped in fast. ------ Quentin says: > An expressive ontology of resilience is harder > because it involves many different institutions, > from local communities to national planning agencies. >... >I think the US effort is developing a relatively >thorough ontology of >response phase communications and subsequently Yes, but, as I understand it (see Newt Gingrich quotes below), it's ITC may rely on things like WiMax relays and perhaps mesh network relays dropped on parachutes which are recovered later. It is actually easier to solve some problems by throwing hardware at them, and Internet resilience may be one of those. This approach may not be appropriate yet for many other places in the world, however. VHF, shortwave, satellite radio, may only be able to move short messages and it's important they be efficient. I'd like to see the ontology be aware of the various paths by which messages are sent during outages, for instance resource requests that compile down to as few radio call letters as possible so they can be repeated over shortwave. That is, someone gives the P-code (or, better GPS coordinates) and some incomprehensible string like "FH5R2X60B1QRQ" which decompresses to "drop us fifty units of plasma and three 12 volt solar power rechargers plus a water purifier before Monday or six people will likely die" when it's fed into software at the receiving end. An advantage of this technique is that it's opaque to the parties relaying the message, so if it's controversial or sensitive or involves moving resources away from a relaying party's own group or region, it's not obvious. And no one can really tell if it's encrypted or not... >resource management, >but resilience will be a much bigger [t]ask. Yes, it's harder, but can it be avoided? William seems to suggest that the problem may not be solved by the same thinking that has failed to solve it before, and when he mentions a National Guard unit "looting" FEMA warehouses, I think of this "ontology of chaos" but I also think of how one manages the transition back to order. Perhaps one thinks of everyone as a "looter" in the first 72 hours after a disaster, and asks them merely to indicate - in exchange for amnesty - what they took from where in order that authorities may repay the owner on some reasonable basis afterward. William also raises the question of unreliable infrastructure and uncertain information, which probably dominates a disaster scenario more than any other kind. It's definitely one of the considerations that differentiates these situations from others. That said, as Paola and Gavin indicate, there are at least a few classes of disaster where infrastructure isn't likely to be affected directly, but can still fail due to personnel outages, sabotage and conflict. Pandemic or conflict scenarios are of this nature. Any AC power grid infrastructure is probably most likely to cease to function, even more than telecom or VHF or AM or FM which are usually required to have backup power precisely because they are needed in emergencies. When they fail it's almost certainly due to relying on AC power anyway. Eventually Internet will be robust also thanks to mesh networks, WiFi and especially WiMax. IPv6 can run on very little power as 802.15.4 proves. When the grid is out, that in itself is a big problem but maybe the way to solve it is to simply assume it doesn't work, and when it happens to work, using it to recharge the batteries, ice and boiled or filtered or pumped water supplies. The batteries may be the worst problem. The sheer quantity of power consumed by field devices can be surprising. A figure I've heard about the US forces in Iraq is that they receive three shipping containers full of batteries for every one of food. Power quality also can matter drastically for medical and communications use. Hospitals in North America use colour coding to designate which plugs are of which quality AC power and which will be kept operating in an emergency scenario. AC power, though, is a luxury in disaster zones, AC inverters eat up at least a third of the power, and for low-voltage devices AC is extremely inefficient. USB provides very high quality 5 volt power and is probably the standard for the future, or 802.3af powered Ethernet. Either way the data and power moves through one plug which is a very major advantage in any crisis situatino. I'd tenatively suggest that three variables are critical in any description of any infrastructure outages or devices/methods to be used during such outages: - DC voltage - many devices run on DC power and use AC only with a transformer that converts and wastes it; If there's a reliable 5 volt (Universal Serial Bus) or 12 volt (car battery) source in a location then you want to know that, if the device will run on it directly then you want to know that. Most electronic batteries today are still nine volt, and hand tools often run on 18 volt, but expect both to recharge from the laptop or the truck in future - many cell phones already charge from 5 volts USB and you may see it in your house soon. - Internet connectivity - can you get an IP number or not? If so, how sporadically? Finding the working WiFi/mesh/WiMax hot spots may allow reconnection of the Internet much faster if relays can be moved to the intersection points between those hot spots and routing software adjusted to restore the packet flow. IPv6 has excellent provisions for doing this sort of thing, and there are standards for very low power IP (IEEE 802.15.4) intended for home automation but that may have applications also for, say, refugee camps. In combination with RFID tags, for instance, it may be possible to assign an IPv6 IP number to literally every pair of shoes or food packet and find hoarders and thieves relatively quickly with some spot audits by people walking through the camp. Or, a destination could be assigned a QR code and a corresponding URI and anyone who scanned it could be told exactly what to do to ensure it gets where it's supposed to go. - non-Internet connectivity - the shortwave or VHF or short range FM transceivers which may allow at least the makeshift coded packets I suggest to be sent, in one of several forms. You could for instance encode them from a handbook even if you had no working cell phone. Many multi-band transceivers now are quite programmable and run Java as readily as cell phones so it's not inconceivable to execute the encoding or even relay IP packets or use digital radio standards. I suspect it's now rational to think of everything as a subset of IPv6 connectivity, since it handles everything from light switches to terabit cables to storage areas networks now. In other words, think of the shortwave communication as effectively a way to construct an IPv6 packet for something to decode. The more you can get into one packet, the better, it can just be resent and resent until it gets through. Packets can even go by carrier pigeon (don't laugh, some microchip companies used to send their designs to fabrication this way in the 1980s as it was faster than any digital method they had available). So, when someone gets a carrier pigeon packet or a shortwave transmission, they enter it in a cell phone or web form, and it all gets eventually to the same transaction system. Since cell phones now have GPS capabilities built in one can at least tell where the message was received from. This would deal well with the situation William here describes: --- William Waites <ww@groovy.net> wrote: > On Wed, Jun 20, 2007 at 11:56:28AM +0700, > In the wake of the hurricanes, the > *entire* communications > infrastructure in the region was destroyed. The > first things to come on > line were VHF repeaters, and as well a > (repeaterless) HF (shortwave) > net was used for early coordination. Several weeks My suggestion is that the power required for these devices and the protocols required to relay succinct requests for, and offers of, help in the absence of more conventional means to exchange them, is paramount. Without that, none of the other eight "clusters" work. > later some cell sites > were re-constructed but they quickly became > overwhealmed (except perhaps > for some prioritized traffic). Another good reason to use programs on cell phones to pass on encoded requests for help (which may even be invisible to the user) is that it takes far less bandwidth than panicky voices. If people realize that an encoded digital request is going to get through much more reliably than voice which a human has to listen to, they will quickly learn to use that UI. At least, those people handling a lot of other peoples' problems (nursing home attendants, doctors, religious authorities, social workers, etc.) can be taught these protocols in advance or trained in them on the fly. It works on receipt, too. Cell phones have cameras on them now, which can read barcodes or QR codes (which carry a lot of information), and should be able to receive RFID information also. If supplies go astray it can be discovered quickly, and perhaps using the RF capability they could be discovered. For instance, if something ends up under the seat of a delivery truck it can be detected as missing and found before the truck leaves. So, I'm suggesting, these request and offer and acceptance and receipt packets are the highest priority traffic, no matter from where they originate. You can filter them for credibility later. You can account for any abuse of the system later. You can prioritize the ones that come from the most credible phone numbers. You can ignore how they got to you, if they can be verified or checksummed or signed somehow. Paoloa suggests also that the requests and offers need not be dealt with only by some group of specialists: > > Assume an emergency takes place where I am (I > seem to have earthquale > > follow me everywhere I go these days). My > friends everywhere else in the > > world will still be able to help > > me better and faster if they can pull up > relevantly filtered information > > using normal search engines and browsers. Correct. They can even relay offers in case there is some doubt as to whether they have what is required - a negotiation could actually be conducted in perhaps three to seven packets total. Perhaps even between instances of say Sahana. Perhaps even between local entities that have already received aid and are now distributing it. Once developed, the protocol can be used by anyone. One can think of the whole disaster zone as a sort of freecycle or eBay perhaps. That's actually what Newt Gingrich says in his May 2006 white paper on emergency response: http://www.newt.org/backpage.asp?art=3081 "A citizen centered system will inherently be permissive, coordinating and flexible. It will consistently seek reasons to say “yes, if” rather than “no because” when approached by volunteers. To a large degree this will be a self organizing system using information technology, expert systems, and citizen initiative, creativity and goodwill to identify and solve problems and develop opportunities at a rate faster than any industrial-bureaucratic system could possibly keep up. In some ways the design of this new system (deliberately NOT an ‘administration”) will be a cross between E-Bay, Craig’s List, and an online dating service. Those who need help will be signaling to those who have help to give and they will be mutually interacting with minimum bureaucratic interference." > Yes, assuming some visibility to the scenario on the > ground. After > Katrina, the problem was compounded by a complete > lack of information. > The only algorithm that worked was to send people > out into the field > with radios, gps (the street signs were, for > example, gone...) and some > note taking facilities (i.e. paper and pencils) and > to report back with > what they could see. Gingrich addresses this too: "Since the 21st century system is dependent on communications and information flow there should be a system for the immediate establishment of wireless cell phones and wireless high speed internet in any disaster area. A combination of aerostats, unmanned vehicles with repeater systems, mobile easily installed ground based repeater systems and other relatively low cost but high leverage systems should be stockpiled (or outsourced under contract for immediate response in a CRAF—civilian reserve air fleet-model). After 9/11 and Katrina there is no excuse for the United States ever again to have a zone of non communications or to have the wireless systems collapse during the crisis." > Of course what was lacking was (1) a coordinated way > of sending out such > teams (this was mostly done ad-hoc) and (2) a way to > receive, categorize > and sort this information. I doubt that the UN terms Soenke provides are fine-grained enough for a world of GPS, wireless Internet and transaction and logistics systems very much more sophisticated than anything most governments now use. Gingrich notes that logistics systems in use now in modern corporations are quite capable of handling an anticipated surge in demand for certain supplies. It makes little sense to invent such systems specific to humanitarian aid when it's probably much easier and more effective to 'borrow Wal-Mart's', so to speak, and just plug in different types of supplies, locations and transport (donkeys carrying medical supplies up mountain paths to refugee camps rather than trucks carrying Christmas toys to malls). This outsourcing may be more straightforward in the more consuming countries, but the capacity to outsource to competent locals in transport and logistics and supply areas exists anywhere. The online request-matching approach, which Gingrich describes as being like a "dating service", lends itself particularly well to using small suppliers of transport, storage and distribution. If nothing else, small shipments that are easy to track can be used to test the reliability and timing variations of different shippers and modes of transport. Obviously, many such small suppliers are doing these jobs anyway, just on an unauthorized or ad hoc basis. As this example shows: > ... > This resulted in, for example, the National Guard > looting FEMA > warehouses for medical supplies to give to community > clinics (set up by > volunteers in the effective absence of the ARC). > > Of course communicating back to the people on the > ground in this > scenario is diffuclt as well... Reporting "we have looted a warehouse to provide the following medical supplies to the following clinics" is just another transaction report. Saves paperwork if FEMA knows its warehouse is empty as soon as it is. A lot of what's going to happen with this ontology is just such status and event reporting after the fact. Gingrich is absolutely right to emphasize the speed and responsiveness of something run by the citizens. Including the US National Guard, who are part-time soldiers who serve reserves and regular troops in a logistics capacity in wartime. It's a logical group of people to train in any advanced logistics methods or codes, not least because there are usually a few of them in any affected area. I don't consider the warehouse incident even unusual, frankly. I think optimizing the ontology to enable just such initiatives and reporting them may be wise. An ontology that assumes top-down management structure for more than the bulk shipping of donated aid items into ports (including airports) nearest affected areas may be missing the point. It's enabling the peer transactions and opportunistic matches between requests and offers of help that will really make the difference. As Paolo suggests when she emphasizes the role of the open Internet and search engines rather than any specialized protocols. To me those specialized protocols exist only to get the message out to the Internet, and allow response from it, not to set up a sort of parallel emergency-specific system. The ultimate achievement would be to relay requests for help directly to forums like craigslist and eBay and job services, and accept non-cash offers of all kinds. A reverse bidding system where donors offer to fill needs as completely as possible might be developed if the supply chain becomes well enough understood. There are some services already to match charity volunteers, charity supplies, but I'm not sure how well they work. Anyone have an opinion of those? ____________________________________________________________________________________ Pinpoint customers who are looking for what you sell. http://searchmarketing.yahoo.com/
Received on Thursday, 21 June 2007 09:35:52 UTC