Yet another software expert explains why the banking system will be hosed by Y2K

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

Subject:Re: mr.milne
Date:1999/08/18
Author:John Denver <jd@howdyfolks.org>
  Posting History Post Reply

Joseph E. McIsaac wrote:
>
>
> Personally, I have been working with banks for the last few years on their Y2K
> programs and also with the Federal bank regulatory agencies some.  The facts and
> opinions "of substance" that I brought to this party did little to sway doomsters
> from their positions.  I stated what was going on, doomsters decried it all,
> calling it spin and me an industry shill.  So much for substance!
>
> Your cry for substance rings hollow -- doomsters believe what they want to
> believe. It's a waste of time trying to change that.
 
I'm not a doomster in the sense you use the word. I have no emotional commitment to the idea that TEOTWAWKI is inevitable. The system might weather the whole thing splendidly, for all I know. I am simply
interested in your reasoning.
 
IMO, the issue that you are sidestepping is the question of interfaces and behavior of the system as a whole. That is one important difference between GPS, fiscal year roll-overs and the actual main event. It is a difference you Pollies need to carefully consider if you plan to keep saying "Jan. 1, 2000 won't be much different than Jan. 1, 1999" and flogging the doomers for their failed predictions. The doomers may have been wrong about Jan. 1, 1999, but they might be right about Jan. 1, 2000 because the latter is QUALITATIVELY DIFFERENT than the former. In all the precursor events, we were dealing with isolated systems. For example, fiscal year calculations are necessarily company-specific because different companies use different fiscal years.
Furthermore, your comments on the soundness of the banking system primarily concern individual institutions, not the system as a whole. This is comparable to the naive claim that a program will definitely work because you have checked all the subroutines. The fact that Company X1, X2... are all "Y2K ready" does not imply that the massive logistical machine formed by combining them is "Y2K ready". That fact can only be tested by going live, and that will only happen on Jan. 1, 2000. (Please don't bother telling me about "large-scale" end-to-end testing. No matter how large scale the test is, it still exercises only the smallest local part of the overall global machine.)
The fact is, we have case studies of large routing systems breaking down in bizarre and frightening ways. At the Denver International Airport, isolated parts of the system had been tested, and when the system was switched on, everyone expected it to perform basically according to spec. In fact, they actually called in reporters, who witnessed and photographed the initial convulsions of the system. That's how confident they were. These were engineers who knew every part of the beast in intimate detail. They were sure it would work. (They certainly knew vastly more about the details of their system, than you know about the global banking system.)
Abstractly considered, there isn't that big of a difference between the DIA baggage system and the world banking system. The DIA system routed bags, and the world banking system routes money. The primary difference is that the banking system is far more complex. So what makes you so sure the banking system won't freak due to a poorly understood
phenomenon like congestion? The DIA system freaked due to
congestion-related issues. As did the MCI frame relay network last week. I'll grant you one point: If I had any money in the bank, I wouldn't be that worried about it getting deleted or lost by the bank in-house. I think you're sincere about what you have experienced, and they probably have done a fairly good job or revising their internal routines. What I am worried about is the banking system seizing up and going down due to problems in the simple routing of transactions -- congestion,
line-balancing problems, queue overflow, inability of transactions to clear, network gridlock and so on.
These are not problems where an engineer digs into the code for *a* bug and finds it. These are systemic problems which fall roughly under the domain of an arcane branch of applied mathematics called operations research (OR). This is why I say your forecasts lack substance. You're apparently not even *aware* of issues like these.
It is clear that virtually every large computer using organization in the world will still have plenty of errors in their code come Jan. 1. This situation seems very similar to the original roll-out of the DIA baggage system. The difference is that the bugs in the DIA system were created in the original implementation process rather than through renovation or rushed replacement.
I agree with you that dependence on technology may be overestimated by the doomers. The system may crash big-time without directly harming that many people. People will find ways to get by, even in the event of massive system failure.
However, I do not find your appeals to "the folks repairing and supporting systems" very convincing. Last week, MCI deployed *teams* of engineers from Lucent, MCI and Bell Labs to solve their congestion problem. Here we're talking *teams* of the creme de la creme, and after 10 days they were still scratching their heads. (I have yet to here even one Polly comment on this in csy2k.) Just as at DIA, I'm sure that the MCI people were hindered in their diagnosis efforts by the size of their system. It's one thing to debug a program on your PC, and quite another to debug a congested frame relay system. Trying to monitor its operation in real-time is a logistical nightmare.
Public disclosures are flimsy at best. For one, there probably is a fair amount of lying going on. And for another, NO ONE is offering assurances that the machine as a whole will not fail due to
systemic/interface threats like congestion, and THAT is what I'm worried about!
 
Looking forward to your comments...



-- a (a@a.a), August 19, 1999

Answers

Thanks a!

I saw this was an answer to J.E.M. and I almost threw up.

"I stated what was going on, doomsters decried it all, calling it spin and me an industry shill."

He IS a freaking industry shill and an arrogant ass at that! He'll do a Kosky and not even address any issue that John Denver raises. I'll make that a prediction and put a guarantee on it : )

Mike

==================================================================

-- Michael Taylor (mtdesign3@aol.com), August 19, 1999.


I agree with you that dependence on technology may be overestimated by the doomers. The system may crash big-time without directly harming that many people. People will find ways to get by, even in the event of massive system failure.

Why aren't we heavily dependent on technology? Do we all live on self sufficient farms? These systems better not be infrastructure related. Even if they aren't, business failures alone will bad enough. What if your company goes out of business and you're unable to find another job (for whatever reason). What's the work around for this? At least if you had some food stocked up, you could eat. But nooooo, this would put you in the doom and gloom crowd. "Don't do that, you doomer", the polly's say.

-- Larry (cobol.programmer@usa.net), August 19, 1999.


And tell me exactly why this tells us that banks will all fail? A whole crapload of rambling about how remediation doesn't work and MCI problems. So I guess putting one or two failures together from the entire world equals certain bank failures.

Still one must wonder how the world functions and why did Y2K cause any problems in just 1999. Personally I never recall a business or gov't agency once having something not run smoothly. Can anyone provide some form of insight into this strange occurence of problems?

At one point in my employment history, I vaguely remember at a video store that a computer system went down. Which then forced our video competitors down the street to close because of computer problems. And then James and Taylor(video supplier) crashed because of us. And then don't get me started on the instant riots that broke out because a psychic predicted in September of 1997 that "A business will have a problem, stock up on tp and get your guns to shoot everyone."

-- MrWayCool (orjustrambleon@forever.com), August 19, 1999.


MrWayCool: Hey, that was ME! Yes, I was the RIOT, I did all the SHOOTING!! Next time, dude, you had BETTER have a contingency plan that WORKS, because NOBODY BUT NOBODY tells me that I can't rent a juicy lesbo mudwrestling video because "the computer is down" or "the power is out" or "the banking system has collapsed" or "martial law has been declared". I don't give a rats ass -- you people provide an essential service to the community, you make sure that you can honor it.

I'm not going to fool around, next time. You have been warned.

-- King of Spain (madrid@aol.com), August 19, 1999.

". Last week, MCI deployed *teams* of engineers from Lucent, MCI and Bell Labs to solve their congestion problem."

Oh sure, this person knows what they are talking about... "congestion" problem? NOT. But it works if they say it was a congestion problem and then tell how "all the banks will have "this" (made-up) congestion problem.

This shows how desperate some people are to prove their conclusions. Too bad they have to swist the facts and make up lies for their "proof". It's "stories" like this that hide the real problems because people are not going to believe anything they hear.

-- Cherri (sams@brigadoon.com), August 19, 1999.



uh, Cherri, it is my understanding that the software was installed without incident and worked normally for a couple weeks. It was only when the system became stressed that the problem was evident and shutdown the works. And I think another word for frame relay network contention is congestion. Oh and BTW, I hear the tried knocking the crap out the mainframe but to no avail... :)

MrWC: We're not talking about mom&pop video rental stores that use a DOS database program. We're talking about the most interconnected industry on the planet. Get a grip. Do your homework. Buy some beans.

-- a (a@a.a), August 19, 1999.


a

Why is it that someone who agrees with you on the banking system is referred to as an 'expert' and the ones who don't agree with you aren't?

Fair question. Would appreciate an honest answer.

Deano

-- Deano (deano@luvthebeach.com), August 19, 1999.


Deano: Denver is a disinterested party. McIsaac is a PR flak. Denver relfects my exact technical concerns, while McIsaac simply glosses over everything. I therefore find it natural to agree with someone like Denver and to discount "experts" like JEM.

I'll admit that I am pessimistic about y2k. As it is potentially a life threatening situation for all of us, there is very good reason to be so.

Honest answer.

-- a (a@a.a), August 19, 1999.


a

Thanks! I appreciate your honesty.

Deano

-- Deano (deano@luvthebeach.com), August 19, 1999.


BEANO!

Do a search in deja on McIsaac - the guy is a raving banking shill of the worst kind, a tecnical know-nothing and an apologist for his crooked scumbag employers...

do you homework laddie...

-- Andy (2000EOD@prodigy.net), August 19, 1999.



Yeah, Deano, then after you do that (if you can remember all of this), read tomorrow's paper regarding the Navy report and how U.S. cities are going to be toast. Then buy some more food for your kitchen. And water, lots of water.

Never mind, you will never remember all this. JUST BUY MORE FOOD AND WATER. DO IT. IMPORTANT. YEAH, BEFORE YOU HIT THE BEACH. RIGHT. Gawd.

-- King of Spain (madrid@aol.com), August 19, 1999.

Moderation questions? read the FAQ