Computer journal compares imported data problem to "Ice-9"

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

In the Y2K world, the Ice-9 equivalent is the Y2K bug imported from outside your organization

-- a (a@a.a), August 10, 1999

Answers

Will you be an importer of Y2K problems?

By John Gantz

08/02/99

In Kurt Vonnegut's book Cat's Cradle, the story revolves around a substance called Ice-9, a form of water that freezes at room temperature. By its nature, if it gets loose from the test tube, it will spread -- to puddles, to ponds, to rivers and finally oceans, until all the water on the planet is locked up in crystalline form. It's an end-of-the-world scenario.

What got me thinking about Ice-9 is some preliminary data I've seen from a recent International Data Corp. (IDC) survey of 1,000 North American companies on their Y2K preparations. It looks pretty good. As of March, less than 4% of companieshadn't started their Y2K work, and less than 2% expected to miss the Jan. 1 deadline. About a third of the companies had already finished testing or were Y2K-compliant.

Enter Ice-9. In the Y2K world, the Ice-9 equivalent is the Y2K bug imported from outside your organization. Your company may be Y2K-compliant, but what about the companies with which you share data or applications? Over half the surveyed companies shared data with others, with companies in banking, finance, transportation, communications and health care particularly wired. Those aren't the industries we want to go down on Jan. 1.

There is actually a quantification of the likelihood of importing Y2K problems from the outside world. Called the Beach-Oleson Pain Index, after its authors, Gary Beach (publisher of CIO magazine and former Computerworld president) and Tom Oleson (IDC's Y2K expert), it relies on the number of external connections as well as the number and type of applications connected. Large companies with complex systems environments, thousands of applications and hundreds of suppliers and business partners have the highest probability of importing problems. (They're also probably spending the most to prevent Y2K problems.)

In the latest survey, the probability of the average company importing a "catastrophic" Y2K problem is less than 0.2%. And even the probability of importing a "business-critical" bug is under 5%. Most of our imported problems will simply be annoying or embarrassing.

Those are pretty good odds. But there's one area the Beach-Oleson Pain Index doesn't address, and that's the geography of those external connections.

From all accounts, Y2K compliance around the world varies tremendously by country. In another IDC survey of 15,000 companies in 15 countries conducted in December, only 8% of respondents in developed countries expected to miss the Jan. 1 deadline with their Y2K fixes. That number was 61% in emerging countries. Clearly, if you are sharing an application or getting a critical data feed from a power plant in Russia or a mining company in Micronesia, your "pain index" may be much higher than the average.

Alas, the first survey I mentioned indicates that one of the last areas for Y2K testing is the data feeds from outside the company. Less than a third have one for testing external data feeds.

Ice-9 was fiction. Y2K problems brought in from the outside aren't. Put some time into preventing them.

wJOHN GANTZ is a senior vice president at International Data Corp. in Framingham, Mass. Contact him at jgantz@idcresearch.com.

-- mabel (mabel_louise@yahoo.com), August 10, 1999.


That book still gives me the shivers! (okay, pun intended) It has to be the scariest I've ever read.

-- CD (CDOKeefe@aol.com), August 10, 1999.

Real-time transactions could be hosed. That means ATM transactions, credit card purchase approval, hotel and air transport reservations, instant "background checks", pigs getting an instant report on your drivers license and car registration (we can hope), etc.

Delayed and/or batch processing could be fixed if imported data is examined (quarantiened) before use. I once worked on a rudimentary "expert system" program to examine data received on tapes before it was used in the system (for an HMO). Of course, this means some programmer(s) will have to develop expert system program(s) specific to each business' needs.

This type of thing could also be done with realtime systems. But has it been done? That (examining input data from others) would be just as much a part of remediation as getting their own programs (and hardware) Y2Kcompliant

-- vbProg (vbProg@MicroSoftAndIntelSuck.com), August 11, 1999.


I read the title as In Circuit Emulator - 9. Guess that pins me.

-- Paul Davis (davisp1953@yahoo.com), August 11, 1999.

Moderation questions? read the FAQ