The legacy of Y2K

The other day I came across a New Yorker article about the time Russia and America Coöperated to Avert a Y2K Apocalypse. The piece recalls how dark some of those Y2K scenarios were:

When January 1, 2000, rolled around, computers all over the world, from stock-market systems and A.T.M.s to nuclear power stations and gas pumps, would jump back to “00,” a catastrophic resetting of machine time that, it was feared, could trip them into failure or malfunction. Airplanes could turn off midair and fall from the sky; financial systems might freeze; municipal water-filtration plants could fail, polluting drinking supplies for millions; and electrical plants might shut down, plunging civilization into darkness. What’s more, the effects of Y2K would persist. Computers wouldn’t simply start working again when the clock read “01.” Experts feared that the breakdown of digital infrastructure could push the modern world into a new dark age.

It makes Covid-19 related restrictions seem minor in comparison.

In 1998, when I entered the software industry, companies in India were in the middle of solving this supposedly monumental problem. Teams in Infosys, TCS, and Wipro were engaged in a number of Y2K projects around the world. I avoided these (software services) companies during campus recruitment. I wanted to work on products, not projects — this is what I told myself.

My assumption was naive: product companies were not spared work on the Y2K problem. While I was lucky to be assigned to a specific product, my wife — who had joined the same product software firm — found herself assigned part-time to a central team tasked with scanning all product codebases and database schemas for vulnerabilities. Her team was on call duty on the night of 31st December 1999, and I joined her in the office campus after dinner. (This was 1999 — ‘work from home’ was a concept reserved for vision presentations on the future of work.) The team watched a couple of movies in a conference room. Around midnight I drifted off to sleep in the ‘bunker room’. The next morning I saw sleepy but smiling faces: there had been zero incidents.

The New Yorker article considers the legacy of Y2K:

“The thing about Y2K,” she told me recently, “was that it illustrated, in very direct terms, how software gets deployed. Pieces of software are black boxes to one another. Different manufacturers. Different companies. Those interfaces are often muddied, misunderstood, and badly documented.” In her view, though the specific risk of Y2K might be over, the broader, systemic risk presented by computer-system interconnection is still very real. “I feel like the lesson has just passed us by, which is the sad thing,” Ullman said. “There’s now a generation or two that really doesn’t know about Y2K, that doesn’t even have any memory of it.”

That fast-fading memory is unfortunate because Y2K times have some important lessons for the increasingly interconnected pieces of software in the cloud today. (Emphasis in text below is mine):

“They all had a sense that things were pretty well in hand,” she said. “But then they slid off into: ‘What about these other people? Can I trust the banks? Can I trust the suppliers?’ The fear of this systemic failure infected people, one after another.” Y2K wasn’t just a technological crisis; it was a social one. It exposed widespread fears of dependence and interconnection that bordered on paranoia. Even if you took precautions, someone else could trigger the end of the world. 

“Fears of dependence and interconnection that bordered on paranoia” — you could say about the current political sentiment across the world.

And you could wonder what would happen if a computer virus wrecks havoc in our highly connected network of software in the cloud. Would that fear of dependence and interconnection return? Would that mean islands on the internet that mirror the isolation nation states are marching towards?

It is now hard to imagine the sort of cooperation between the U.S. and Russia the article talks about. Which is another reason why we need to keep alive the memory of those Y2K years. There are times when the world has to fight a common adversary. We are living in one of those.

Update: Not everyone agrees with this narrative of Y2K. Bob Cringely is one of those with a different view on how interconnected systems were back then:

…when I jumped into the research in 1999 I found that Y2K remediation, as it was called, seemed to be going well. I also found that systems weren’t as inter-connected or dependent as many of us had thought — that the world simply wasn’t as much at risk as we feared. I had to fight for this position, but ultimately that was the more conservative story we told two months before the actual event. And we were right.