Redundancy and 'ifs' re communications - by request : LUSENET : TimeBomb 2000 (Y2000) : One Thread

Now I have been asked just why I do not think everything is going down due to Y2K problems. Why the web of infrastructure will not fail due to interrelated dependencies and things just generally quit. This is not an easy question to answer, from a general overview standpoint, which is the viewpoint that has been requested. When I first tried to put the answer into words, I was not satisfied with the result at all. Explaining this sort of thing to another computer scientist is easy, but he does not need the explanation! And using the terms I was trained in will just convince everyone I am spreading a bunch of BS over real problems. And it was entirely too long. So I am going to address the question of the communications infrastructure first, then power and transportation later  my ideas about power and rail are pretty well known anyhow, but communications is a subject of a higher complexity, and not that easy to address in everyday terms. (Now dont get the idea I am talking down to anyone. People spring a new buzzword on me about once a week, and it annoys me, so I can imagine how non-techies feel about it.)

So, without further ado, here is my explanation of why I dont believe the communications system will fail due to Y2K problems, with lab work yet! :^)

Get a large piece of paper, a pencil and a ruler. Mark about 30 points in a rough circle around the paper  about 6 inches across will do. Then take the ruler and draw 5 to 10 lines from each point to some of the other points. Draw your lines randomly  just make sure everything is connected without trying to make some sort of regular spirograph pattern. When you are done, you will have between 150 and 300 lines drawn on the paper. By tracing the lines, you will see there are many points without a direct connection to each other, but they can all be linked between a second or third point. Now draw a line from one of the points directly out of the circle so you have a sort of dash or tail sticking out on one side. Now do the same on the other side. Label one of them A and one B. Trace a path across the circle from one of the marked points to the other. That is pretty easy. NOW  delete one of the lines you used to trace across the circle and find an alternate route. Try deleting several lines at random. You can probably still find a number of connections that will interlink point A with point B. Fiddle with this for a while  you will soon convince yourself that an awful lot of random lines have to be deleted to completely disconnect any section of this network from any other section. (Sure, you can do it on purpose  but that is cheating.) You may take a dozen hops to get somewhere, but you can still get there. And the sections will still be in contact internally, even if you manage to disconnect some part of the net from another part, they just cant talk to certain places across the breaks. Now you have an idea of why the Russian phone system acts the way it does. ;^)

Now the dots represent routing centers in the communications system. The lines represent trunks  which can be satellite links, microwave relays, fiber optic links or even copper cable. There are a great many more than 30 major routing centers and certainly the average major routing center has more than 5 trunks. A phone call (for example) is usually broken into digital packets, the packets sent separately across the routing network, reassembled on the other end, and sent across the local system to your phone. No rule is set about the route it may follow, just that it must get from point A to point B. A call may well be rerouted several times a minute, and you will never know it. Knocking down this system takes more than a few problems  and more than a few power outages. Even if it broke into several subsystems  communications would still be possible from point to point INSIDE the subsystems.

Now the system described above is the national system. Each point would represent a routing center in an area code. Now imagine (I expect you are tired of drawing lines by now) another diagram of perhaps 15 or 20 points, interconnected as above, with threads tied to 5 or 6 of the points. Run these threads to one of the points on the larger diagram. Now you have diagrammed a local service area of a local phone service provider. One or two threads can break and you still are connected to the national system, even if all the threads break you are still connected locally.

Now if you have done as I requested above, you now know the complexity of the system is not its downfall as some claim. It is its salvation. Perhaps a nuclear war or a major meteor strike might knock out enough of the system to destroy it. I cant imagine anything less as doing more than breaking it into several subsystems. And the routing centers have backup power for emergencies  and a lot of other things have been done to protect various parts from natural disaster. And the govt. imposes certain things to make the system safer and more secure  but its real protection against disaster is simply the complex level of interconnectivity inherent in the system itself.

Whew! People ask me why I am not a teacher. Answer: Its hard work!

Anyway, that is why I dont think the communications system is going down. Let me know what you think. Post elsewhere if you want to, just leave a message here if you do so I know about it.

---------------just a few BTW remarks below-------------------------

A couple of notes here - billing problems can knock off individuals - not the entire system. I can't speak for all the local service providers, but the ones I have hooked up to in KY and TN have always had a human in the termination of service loop. And a human is going to look at an enormous pile of termination notices with a certain skepticism.

A lot of comm equipment is pretty new right now - this is due to the rapid growth of the Internet - and is the proximate cause of the push to set some sort of surcharge for Internet long distance and such. The providers want to get some of that big investment back in a hurry. But new equipment is much less likely to have Y2K bugs - and much in the way of routing equipment and so forth does not care about century dates anyway.

Hardliner - this is what I mean by long chains of if - you must convince me of an awful lot of 'if this line goes down and if that happens' to knock down a system designed in this manner.

Arlin - am still working on redundancies and oversupply in the food chain - there looks to be more than even I thought, but I am trying to nail down better numbers before I post it - I don't like making too many assumptions.

To the Heinlein lovers - yes, I did think about using Heinleins' description of a comm network from 'The Moon is a Harsh Mistress'. But that description, while a somewhat better structural representantion in some ways, is quite complex, and rather difficult to work with. So I went with a flat model instead.

General remark about this type of network - the complexity of a large system is very high. Some unusual things about such networks are known - for example: it is not possible to compute the best route for a message in any reasonable time. A good enough route, easily done, but the shortest route is an extremely hard problem. Many other non-obvious things have been discovered - and many books on the subject are available.

-- Paul Davis (, January 14, 1999


Well folks it looks like we all can pack up our brains and get on with our life. Today I'm calling the food bank and donating all my stash to the poor. I'm saving my stored water to water my garden with this summer. I won't have to buy TP or any of the other essentials for 6 months, I'm glad about that because I'm sick of shopping and putting away this stuff. I am so relieved that Paul has set us straight, I can now have a decent nights sleep, and take that well deserved vacation. I'm glad I won't have to go to the bank for several months to get cash out, I'll filter it back into the system little by little. I can get back to reinvesting my 401K and other investments. I'm mad at myself for swallowing this whole Y2K thing hook-line-and-sinker. The next time someone cries holocaust, I'm going to telling them to go screw themselves! Bye you all, it was fun while it lasted you won't be hearing from me again.

-- navajojoe (, January 14, 1999.

Sure hope you're right. I've no reason to think you're not. I reckon we'll know soon enough.

I don't have the technical background to comment on details. I know that in recent history trivial software errors in telecomm have induced major service interruptions. Soon restored, of course.

Will remediation work itself introduce no undiscovered errors?

Will the notorious embedded systems present no problems?

Will any problems only appear one by one, to be dealt with case by case, in a normal environment?

Could any such problems interact destructively?

Does telecomm require the power grid to operate?

Will the grid be up?

Is the answer to each question maybe?

From where I stand, uncertainty is still king. Or regent at least.

Koskinen and Abrams, Connor of the U.N., Philbin of the National Guard, and others all have indicated some degree of uncertainty themselves.

-- Tom Carey (, January 14, 1999.

Amen Paul and Navajojoe

I'm heading for the hills alright, when I'm 65 and retired from society the hard way by working my way through it.

-- Adam (, January 14, 1999.

Alright, you say that it can't take *everything* out... That is correct. Not at once...

What about secondary failures...? What happens if enough individuals are laid off and can't take that vacation...? What if they don't have to buy any towels anymore due to people not stealing them..? What happens to the people that depend on visitors...? What happens to the company's that make towels..? What if they can't sell enough towels to stay in business..? What happens to the suppliers of cotton..? What if they can't pay their bills...?

One example... It's a chain reaction... Yes we are not *completly* interconnected... BUT.. a few failures can affect everyone...

One power company in idaho fails and takes down 1/4 of the grid...? That can't be possible... I'm dreaming again right..?

In the news: Down in Texas.. Dallas County, to make sure their Childsupport payments could accept y2k a new system was installed. Ooops, new system does not work right... Support payments are now already 2 weeks delayed... Hmnn.. yeah... I am dreaming...

Sorry, just a little cynical this morning...


-- STFrancis (, January 14, 1999.

Paul, Correct - as far it goes. But your thinking as an intelligent, reasoning person; not as a blindly dependent stupid computer.

The computer can't think and instead will do something wierd and wondrful - and totally unpredictable.

For example. Termination. Yes, if a program fails, many thousand (possibly millions) of accounts (possibly all of them) will simultaneously "fail" and may be called for termination. Each needs to be "fixed" by hand - or by re-running the original program once it is fixed. By who? How? If services, power, HVAC, or water is down in which area? If bills (mail) is delivered to one place, will its home bank still be up - no power no check transfer!

What does the billing company do for money in the meantime? How can you suddenly track and fix ten thousand failures a day if previiously you have received ten dozen a day?

Rerouting around failure had to be specifically programmed into the Internet by each university and DARPA as survivability requirements in case of nuclear war. Now, telecoms and FAA traffic control systems do the same thing manually and automatically - if they can communicate withe ach sensor and region. Other grids (heating oil, chemical plants, power, gas, water, sewage) are inflexible - they cannot be re-routed randomly like electronic signals because they require power, sensors, phyical (one-way) connections, and controls that ultimately demand point-point service.

Draw your network, but draw one-way arrows on half the lines. Take out your white out - and randomly whiteout several nodes. Use the whiteout to "draw" a line across the paper - - now try to get "services" across the network. While blindfolded - with somebody else telling you what to do over the telephone when they are looking at the same network upside down.

-- Robert A. Cook, PE (Kennesaw, GA) (, January 14, 1999.

Duped from a earlier thread:

Also: American ingenuity and inventiveness not withstanding - assuming there is some left after a government education - but there is absolutely nothing an "American citizen" can do to get water, 911, phones, satellites, power, or natural gas back after an infrastrucutre failure.

You can only sit at the end of the pipe, waiting for somebody else to recover each and every section.

The situation is worse in distributed systems needing multiple silumtaneous systems to also work: even the expert id held hostage by factors he or she can't influence.

And so the repairman (or programmer) desperately working on the SCADA system to work in Atlanta to get the natural gas flowing through Lousiana can't do anything about the power outage in Houston that is causing the pumps to fail in Beaumont so there is no gas pressure in Mississippi to get gas to AL and GA!

And the satellite relaying its flow information from Mobile to New Orleans to the control pumps north of Lake Pochetrain is now pointed towards Antares, not the sun. Pluto will get the signal -if it's transmitted at all.

And the Atlanta power station that is keeping his lights on has three hours of natural gas supplies left before it trips off line.......

-- Robert A. Cook, PE (Kennesaw, GA) (, January 14, 1999.

Thank you for taking the time Paul, to illustrate so clearly the communications web in a concept easily understandable. Ill use that graphic example. Its a good teaching tool.

In adding to others comments, Im not sure I completely agree will your assesment, however, from the standpoint of all the interdependencies. If there is a brownout in some areas, and blackout in other areas, what does that do to the communications system? If a Y2K terrorist (such as the DoD, National Guard and FEMA, et. al. behind-the-scenes activities and concern indicates is possible) gets involved in impacting the interconnectedness, how does that impact communications? Then theres good ole Mother Nature, seemingly more and more upset each year and tossing new natural disasters our way. Then theres the crumbling economies in Indonesia, China, Japan, Brazil, Russia, et. al. Then again there the U.S. national rush to Y2K compliance, or not. And then theres the United Nations Y2K trickle-down effect. And, and...

In an isolated world your analysis may hold Paul. Were not isolated on this one.


-- Diane J. Squire (, January 14, 1999.

Although I am preparing for TEOTWAWKI and consider that a legitimate technical possibility, from my knowledge of comm networks, I believe Paul is a heck of a lot more right than wrong. At worst, we will have highly significant "islanding" with comm (though losing Brazil, say, as an island, is not a happy thought)

That said, we should be careful about extrapoloating across industries, even highly networked industries. For instance, utilities have an entirely different type of redundancy, which can't be generalized the same way as with comm.

Picture industries as though they are "object classes", each with their own properties, methods, etc. Sure, they communicate through a set of clean prototocols and messages (ha! sometimes), but the classes are, by design and definition, distinct.

Look, I'm not trying to push that analogy too hard. Point is that we absolutely need to "distinguish the things that differ" (so thanks Paul on comm, people need to know that) but it don't change the fundamental picture which is, intense uncertainty and major worldwide risks.

It is Y2K's pervasiveness across all the "classes" and the reintroduction of those "fixed" (???) classes into the world-system more-or-less simultaneously (I view a 12-month period as functionally simultaneous) that brings the unavoidable risk.

It's boring. It's dumb. We didn't start in 1995 and there is no getting around that.

-- BigDog (, January 14, 1999.


I think I'm beginning to understand your perspective a little better. You come across to me more as a "builder" than a "fixer".

It's obvious that you've put time and effort into your explantion and it shows. It is a sound and clear instruction.

From my viewpoint though (that of a "fixer"), it is an explanation of how it all works when it's not broken, or how it is supposed to work. Also, one of the details that I find missing is that any of your myriad paths or routes that require tolls have a requirement beyond simple communication. I am mindful of an AT&T data center that I once worked at which (to illustrate the size of the center) had three full acre sized floors of nothing but DASDI. This place had many, many racks of 3 inch wide yellow paper tape that contained punched holes detailing the times, duration, source and destination of each toll call that went through AT&T!

Now if the time did not include the year, or the calculation of duration did not involve timekeeping in the extreme, there would be no problem with this. I haven't specific knowledge of the code at a low enough level to say, but it's a fair bet that AT&T isn't spending all that boodle because there's not a problem.

By way of anecdotal evidence, there's a guy in my neck of the woods who's retired from the phone company after 28 years worth of fixing stuff with meters, wire strippers and schematic diagrams. During our first conversation about Y2K and the possibilities, his asessment was that since every phone company in America got it's time standard from St Louis, toll circuits would not be available unless somebody got their act together. As best as I can remember his reply was something to the effect of, "Well, you'll get a dial tone and you'll be able to call anyone in "Birdville", but toll calls ain't gonna happen." (in fairness, I do not know the man well and I have no idea what any of the details about "St Louis" are or if this guy really knew what he was talking about)

Now, as I said, I don't have specific knowledge of the code, but I'm sure enough to bet my life that the phone company (which measures your calls in fractions of minutes) is not going to connect you unless and until they are sure that they're able to keep track of how much it's going to cost you. (And I'd bet it again that RLH would agree!)

In short, the robust and redundant communications network that you've described so well is not going to be allowed to function unless they think that they'll get their money.

Can you address this aspect of communications networks?

-- Hardliner (, January 14, 1999.

Paul, I think your point on communications is correct to a point. As long as there is power to a compliant local exchange office, there will be local phone service within that exchange.

Long distance has a better chance to hold up today than if Y2K occurred ten years ago, because of deregulation. More long distance carriers with separate facilities means that one or more links should still operate between any two points, as long as power is available. If Y2K occurred under the phone system of "just AT&T" available for long distance, then it would be high probabilities of failure in long distance.

That said, there are still some pretty strong possibilities of single- point failures taking down long distance and leaving us with disconnected, but still functioning local systems. Use the Los Angeles metro area as a hypothetical example. Suppose every major player in the long distance industry has a switching center somewhere in or near there. A regional, long-duration power outage in the LA Basin would pose a threat to long distance service to much of the west coast if it lasted long enough to exhaust the phone companies' fuel supplies.

Same for the other places where the various networks converge. The international phone system is concentrated into even fewer "choke points", leading to higher possibilities of single point failures taking down the system.

My overall outlook for phone service in a post-y2K scenario is that we are most likely to have local survices survive, but long distance and international services are likely to resemble such services fifty years ago; relatively expensive and rare.


-- Wildweasel (, January 14, 1999.

Paul, thanks for the post. I want to address this quote: "the phone company (which measures your calls in fractions of minutes) is not going to connect you unless and until they are sure that they're able to keep track of how much it's going to cost you. (And I'd bet it again that RLH would agree!) " I know for a fact that the phone company will connect you. Traffic systems will (ours are Y2K compliant now) track all calls. They will also bill you (our billing systems are Y2K compliant now). They connect you today not knowing if you're a good customer or bad customer; that's how these wrap arounds (10-10-XXX) work. They bill you after you make the call. The big companies didn't get that way because of stupid business decisions. They know about Y2K and are fixing it.

Let me address one additional point: the need for power. The key facilities have back up generators. So even if power is down, calls can get through. I remember the San Fancisco earthquake. I couldn't get through to my husband there on business. He managed to get to a pay phone and called me. Power was out, phones were out, banks didn't fail and grocery shelves were still stocked. Amazing maybe we can survive Y2K.

Troll Maria

-- Maria (, January 14, 1999.

Hey, no flames, eh Paul? Flaming's not endemic to the EOWers, I think, it's just that this is the first time you've come to the table with a really well reasoned argument.

The upside to Paul's argument is that it does not presuppose that everything will be fixed; that is Troll Maria's domain and the faulty logic of extrapolation from her own personal experience clouds her vision beyond argument, I think.

I guess, Paul, that I still question whether a certain critical mass of telecom loss, bank failure, and power loss does not result in a domino effect that takes out the rest. For example, power goes off and stays off for an extended time in 25% of the U.S. This overwhelms the backup power of the telecom providers in that area so even if their stuff was fixed, its carrying capacity is lost to the nation. Some 25% of the telecom (some overlapping with the powerless area) nationwide is lost due to Y2K glitches, leaving the U.S. with, perhaps, a 40% degredation in telecom infrastructure.

Ok, now we find that there are two or more Federal Reserve banks and the headquarters of a couple of other major banking institutions in the areas affected by power and/or telecom blackouts. Remove power and telecom to these major entities for more than a few days and the national banking system collapses (anybody care to challenge that supposition?).

The collapse of the banking system makes everybody pretty jittery. Lots of absenteeism at the remaining telecoms as husbands/wives, moms/dads stay home to make sure things are ok. So function of the remaining 60% of the telecom infrastructure begins to degrade. Keep that up for a couple of weeks and perhaps we're down to 30%; as that degradation continues, it has a ongoing degrading effect on the remaining utilities. It becomes more and more difficult to remediate anything because we're always working on backup power, or experiencing brown-outs that crash our systems which then have to be rebooted, or trying to get ahold of ole Troll Maria who decided that she had better stay home to make sure the kids are ok.

I'm not disagreeing with you; I'm not agreeing with you. This is why I think that a very serious collapse is still entirely possible and therefore why I am preparing accordingly. There's a lot of "ifs" there, but it does not strike me as one bit implausible. And that's what has me reacting as I do to Y2K. It's the very first EOW-style scenario that strikes me as completely plausible. That's what bothers me the most. And therefore total, all-out preparation is the only rational response.

-- Franklin Journier (, January 14, 1999.

turning off bold.

-- Franklin Journier (, January 14, 1999.

Help! Did that do it?

-- Franklin Journier (, January 14, 1999.

Geez, sorry folks. Here we go again

Did that do it? If not, I give up.

-- Franklin Journier (, January 14, 1999.

Troll Maria,

You either need to learn how to read a little more carefully or to be honest enough to address the point at issue.

I said nothing about determining whether anyone was a good or bad customer before allowing a call. If you will re-read the post, you will find that I said they would not connect the call unless they could price it. You simply asserted that they could price it now and that they will be able to do so post Y2K. If you're simply trying to communicate your company's readiness, why introduce the unmentioned (except by you) issue of whether or not someone is a "good" or "bad" customer?

Sorry, your tongue still appears to be split and your defensiveness isn't helping any.

-- Hardliner (, January 14, 1999.

bold off.

-- Jack (, January 14, 1999.

lemme try

-- Hardliner (, January 14, 1999.

I really didn't extrapolate anything. You're the one doing the extrapolating. I really don't understand that from a few power outages you get to 40% of telecomm failing then of course that leads to banks failing. You make two very big leaps in your extrapolation. Leaps that I can't seem to make.

But we have real failures today (Asia, Russia) and somehow the earth keeps spinning. Businesses do depend on banking but even a few weeks of outages will not bring down the banks. Bad investments bring down banks as evidence last summer. My main view has always been that you can't base any Y2K scenario (good or bad) on no information. I'll always stated my opinions as opinions and my facts as facts.

Troll Maria

-- Maria (, January 14, 1999.


I expect we're at an impasse. I fully admit that I am extrapolating -- I prefaced my remarks with "For example..." I say that if we lose 25% power for any extended period of time, the telecoms in that area will go down; they don't have enough backup power for an extended outage. Dispute that if you have other evidence.

If another 25% of the telecom network has its own troubles then we're talking about ~30-40% reduction in service nationwide. I'm making those numbers up for the example, but they do not seem implausible to me.

It seems fairly clear to me that the financial system as it is presently structured could not survive a couple of weeks of power outages and loss of telecom at two Federal Reserve banks and several other cornerstone banks. The various financial markets would be crashed, individuals and businesses would have lost most of their personal wealth. After that, who would immediately put their money back into the electronic system, even after the power is restored?

But, as I say, an impasse. It seems entirely plausible to me; that's why I am preparing as I am. It seems entirely implausible to you. That's why you're not.

-- Franklin Journier (, January 14, 1999.

This one's for Hardliner. Sorry I didn't explain Traffic systems. I thought you knew, my faulty assumption. Traffic systems maintain the records of long distance calls. They record call origination, destination, time start and stop, and a bunch of other stuff. If something would go down in our Fraud systems (which detect "bad" customers), traffic will still continue to store these records in their databases. We would still know who called whom and when and for how long. Sorry I went off into the billing system. Didn't mean to confuse you. Both system are compliant and have been future date tested.

Troll Maria

-- Maria (, January 14, 1999.

Pretty encouraging post, Paul. A couple comments:

1) A common cause that brings a lot of failures at once takes out some of the redundancy, and

2) As an Internet programmer, I'll comment on the one network I know anything about: Even though the Internet was theoretically designed to survive nuclear war (at least that's the legend, I've seen it disputed), about a year ago the failure of a single key server took out a quarter of the Net until it was fixed. Also, Net traffic is currently routed through a small number of backbones. It's more hierarchical, not the massively parallel topology you describe.

-- Shimrod (, January 14, 1999.

Troll Maria,

You neither confused me nor addressed my assertion that the connection would not be made unless the carrier was able to price the call.

All you have done is sidestep by saying that your carrier can and will be able to price the call.

Aside from the credibility issue raised by your apparent evasion you still have not said, "Yes, we will make the connection even if we cannot price it", or, "No, we will not make the connection if we cannot price it." Neither have you addressed whether any carrier would do so.

I've got to tell you, you sound too much like the telephone solicitors who read from a list and simply parrot the desired message instead of answering the question asked for me to put much faith in your words.

-- Hardliner (, January 14, 1999.

Hardliner wrote: "the phone company (which measures your calls in fractions of minutes) is not going to connect you unless and until they are sure that they're able to keep track of how much it's going to cost you" I agree with the validity of this statement. Here are words you'll understand: They are sure that they can keep track of how much it's going to cost you. (Traffic and billing systems) Ergo they will connect you. Your statement was a simple if p then q statement. I verified p is true then q will happen.

Troll Maria

-- Maria (, January 14, 1999.

Troll Maria,

You should try selling used cars! You just won't deviate from the party line will you?

You seriously patronize the participants in this forum by presuming to know what words they will or will not understand.

I didn't hear anyone say that they didn't understand the words. What I did hear was you saying everything BUT, "Of course they will not connect you if they can't figure out how much to charge you!"

By the time one is ten years old in America, it is clear that the "phone company" will send a bill to the executor of your estate if they know his name! OF COURSE they will connect you if they can price the call. OF COURSE they will let you run up an enormous bill whether or not they think you will pay it! OF COURSE the Yellow Pages are yellow! None of that is the point at issue.

I could not care less about your evaluation of the "validity" of the statement. Can you say, "That is a true and accurate statement without choking (the statement that you have now twice quoted)?

-- Hardliner (, January 14, 1999.

Hi folks. Glad to see so many of you read and understood my post. When regarding Y2K problems in the comm net, it helps to remember the Moscow phone system. Ancient equipment, bad wiring - I honestly think the system would be improved if they used all the old stuff BellSouth and AT&T sold or auctioned last year. (Get to the sales from their websites under surplus - auctions aren't on the web yet or not last time I looked) But the system still works for most of the people most of the time. Just frustrates the crap out of anyone used to something better. Now my idea of the very worst we could be reduced to would be the Moscow phone system. And it would not be a permanent problem - we do have the fiber optics and so forth already in place.

The Internet backbone does not have quite the redundancy of the phone system yet - I think that is changing as more demand for bandwidth requires expansion. One of the points I glossed over in the example was that adding a routing center increases bandwidth greatly beyond the simple arithmetic increase you would expect - as it adds so many other possible routing links to the other routing centers. Maybe in another post sometime.

Hardliner - the comm companies don't give details of their billing system to outsiders - and since I am not a phone Phreak I have not tried to hack the system. I do know the Phreaks don't seem to have much trouble bypassing the system - so I would presume the billing requirements could be turned off during an emergency without much trouble - say after a Presidential Executive Order or an order from FEMA - with compensation to come from the general fund after the emergency is over.

I do know the routing centers have emergency power - they don't advertise how long it will last using only what is on site - I know it is longer than three days - because I have been in a wide area blackout for three days and had phone service the entire time. There might be some info from the Public Service Commission on this. BTW - have you ever seen the ad they run down here with the guy who gets locked into a routing center during an emergency drill? He panics when he finds out the doors will be locked for more than a week! Don't know if that is true and they have drills that last that long, or just Madison Avenue talk - cute ad though.

-- Paul Davis (, January 14, 1999.


Your presumption that the billing code could be bypassed fairly easily seems quite reasonable (at least in a technical sense. The "business case" might be a harder sell) and if an emergency put enough at risk, an EO or FEMA directive is a likely expectation.

I suspect though, that rather than compensation from the general fund (or any compensation at all) the order simply would ban toll calls to all except government and "approved" entities. Somehow the idea of Uncle Sam coughing up to pay for the population's long distance bill doesn't quite ring true. . .

As for the rest, I have to tell you that your information is "hands down" the single best piece of evidence I've seen that we are unlikely to lose the entire phone system, at least initially or all at once.

I look forward to your views on power and transportation and whatever else you might present in such an excellent manner.

Thank you.

(BTW, I have a semi-permanent "brain check" that erratically and erroneously identifies "The Great One" as RLH instead of RAH. Apologies to all RAH fans!)

-- Hardliner (, January 15, 1999.

In reading through the comments on the thread, a lot of people are quite rightly wondering what the effect of other layers of interdependent connections would have on the model that Paul Davis has proposed for communications (e.g., the collapse of the banking system or loss of the power grid). In following Paul's approach towards the circle with the dots that randomly interconnect, it seems to me that to account for this, one should simply add more dots to represent these other entities. I mean, why limit yourself to communications, when you should be able to define an interconnected, dependent structure that also reflects how the banking system affects it, etc. In fact, it is more meaningful to consider where all these random lines intersect as also critical points -- so, for instance, two intersecting lines at a point might have four communications nodes sharing a power utility as their common link -- if the power utility can't deliver power, then all four communications nodes are hurting.

And if you do this, you start thinking of Paul's model of How Things Work more like a web. And if this approach sounds somehow vaguely familiar, be advised that it should. Paul's model is nothing more than a more restricted (communications only) version of

Charlotte's Web

offered months ago by that notorious doom-and-gloomer, Infomagic.

That having been noted, it really just becomes a matter of judgment as to "how many" strands of the web can fail before the entire thing collapses (Infomagic's de-evolutionary spiral).

My own view is to imagine Paul's circle as a frying pan, with Infomagic's (Charlotte's) web stretched to the rim. Then:

SCENE 1. [Image of circular frying pan with spider web.] "OK, one more time. This is your world."

SCENE 2. [Egg is broken over pan. ] "This is your world with Y2K."

SCENE 3. [Egg sizzles over the web, causing it to completely collapse and burn.] "Any questions?"

-- Jack (, January 15, 1999.


You and I have disagreed in most of our posts. But on this issue you at least have a reasoned premise. I would like to encourage you to continue in this new direction.

I do have one question about your scenario: What would be the impact of dramatically increased traffic on the the surviving routing centers? How much of the slack in the system would they be able to absorb? At what point would the increase overwhelm them?

As my question probably shows, I'm not an expert in this area; but I do see potential problems because of overloads. Thoughts, anyone?

-- Nabi Davidson (, January 15, 1999.

Moderation questions? read the FAQ