Those Year 2000 disaster preparedness trucks were originally there for a different reason


I mentioned some time ago the trucks filled with computers and backup generators parked outside Building 26 on the evening of December 31, 1999, standing by to assist with operations should a worldwide disaster strike during the change to the year 2000.

(Spoiler: Nothing major happened.)

A now-retired colleague of mine¹ told me that those trucks were originally there for a different reason.

Those trucks with generators had been parked outside Building 26 for months before the end of 1999. They were originally there for reliability testing.

The reliability goals for Windows 2000 included continuous uptime that exceeded the time between power failures in the Redmond area. Although the Pacific Northwest is rich in hydroelectric power,² we also suffer from a lot of power outages during the windy autumn and winter months. That's what happens you have a lot of tall trees, I guess.

The original plan was for the trucks to be returned when the testing was completed in December, but upper management decided to extend the lease through to New Years, just in case.

A lot of critical infrastructure needed to be updated with a stable operating system well before the end of 1999. The software that runs banks and hospitals are closely-regulated, and a high degree of continuous reliability is one of the criteria. Customers in those industries (as is typical of most industries) have a lot of third-party software that needs to be tested. This software is highly specialized and won't exist in Microsoft's application compatibility library.

Enter the trucks with computers and backup generators. These systems were set up to replicate the necessary computing environments and set to task running tests, including so-called "long-haul" testing, which keeps a system running for extended periods to help flush out issues like a tiny memory leak that eventually grows to the point where the server becomes unresponsive.

My colleague notes, "There were a lot of extra hours put in to make sure the industry was supported for an event that would happen only once."

¹ In the years leading up to his retirement, my colleague was, among other things, responsible for maintaining Notepad. He wasn't the original author (the identity of whom has been lost to the mists of time), but he was the one in charge of keeping it running.

² Although coal also plays a major role.

Comments (13)
  1. DWalker07 says:

    Well, Year 2000 will happen only once, but Year 10000 will be here “soon”.

    1. smf says:

      Only 21 years to replace old unix systems.

      https://en.wikipedia.org/wiki/Year_2038_problem

  2. Withad says:

    I think you actually solved the mystery of the Notepad author about a week after that first post – https://blogs.msdn.microsoft.com/oldnewthing/20091117-01/?p=15993.

    (Wikipedia attributes the very first version to Richard Brodie but that doesn’t seem to be who you’re talking about, so perhaps there’s still some mystery.)

  3. Ian says:

    Although the Year 2000 event was a one-off and passed without any of the predicted drama, another will occur on Tuesday, 19 January 2038 when signed 32-bit integer time will overflow. Although all major OSs will have been using 64-bit integer time for a several years there will still be plenty of embedded systems that don’t and could roll back to 13 December 1901. I wonder if this time it will go as smoothly, or if complacency will have set in. Or perhaps the number of affected systems will be vastly greater than last time, they will be less accessible, and identifying and fixing them will prove an impossible task.

    1. Zan Lynx' says:

      A lot of “cloud powered” IoT devices will fail in 2038 because they use 32-bit systems and SSL/TLS libraries. When they check certificate expiration they will fail. When they DHCP and / or get network time, they will fail.

      Whether or not any of today’s Linux powered light bulbs are still running in 2038 is a different question. I doubt it. I don’t think the vendors will bother maintaining the domain names and web services.

  4. Martin Ibert says:

    The problem is not the tall trees. The problem is that in the US, power lines are not usually buried as there are in many other countries, rendering falling trees harmless to the power infrastructure.
    Backhoes, on the other hand …

    1. Yukkuri says:

      Utilities generally have found buried power lines to be LESS reliable than lines on poles, not more. In fact you have to over-provision them to compensate for the fact that when they fail you have to dig up a street to repair them, which is just not something you can do many times a year.

      Those small boxes you see on the streets in neighborhoods with underground wiring that have dire warnings about electrocution are points where they can swap around between redundant cables when one fails to quickly get service restored.

      1. Marcel says:

        Well, whatever the case, I honestly can’t even remember the last power outage here in Germany, certainly none within the last 3 years. And we do bury our cables. On the other hand electricity is at least twice as expensive as it’s in the US, so maybe it’s just a case of “you get what you pay for” ;)

      2. When an underground cable failed in our town we (and 50 or so other properties) were fed from a large diesel generator, deposited on a nearby verge, for several days before a repair could be effected. Admittedly the temporary supply was 100% reliable, although I expect the frequency was off.

      3. smf says:

        Most of the power cables in the UK are underground, although the power is distributed around the country with cables on pylons. Powercuts are usually caused by one of those going down, or a fault at the substation.

        We don’t get much seismic activity, which may explain why the cables are safer underground than above.

    2. pc says:

      Putting power lines underground is just too expensive. It may work in some parts of the country, but where I live in Massachusetts the ground is full of rocks and ledge. There are serious estimates that it would cost a trillion dollars (yes, with a T) to bury the lines in just Massachusetts.

      Here’s one source: http://boston.cbslocal.com/2011/10/31/patrick-1-trillion-to-bury-power-lines-underground/
      And I know I’ve heard “it would cost trillions” in person from some National Grid spokesman at some point on the issue, but I don’t have an easily citable source for that.

      1. morlamweb says:

        And trimming trees that grow up near/around power lines costs far less than burying the lines. There’s been a big push by utility companies in the Northeast U.S. to trim vegetation back from the power lines after several crippling storms came through in recent years.

  5. Neil says:

    While obviously someone has to be responsible for maintaining Notepad (otherwise you would have to remove it), it seems more of a symbolic appointment; I can’t imagine there’s much actual maintenance.

Comments are closed.

Skip to main content