Good Morning! Nice of you to drop by. Today is Thursday, Nov. 20, 2025.
Today in History:
1998: The first module of the International Space Station is launched.
1985: Microsoft releases Windows 1.0, giving rise to the axiom about never buying version X.0 of anything at all.
1945: Nuremberg trials begin.
1923: Traffic signal patented.
Birthdays include Joe Biden, Robert Kennedy, and Crown Prince of Austria Otto von Habsburg.
=-=-=-=
Let’s talk today about the outage at Cloudfare the other day, and the Amazon Web Services outage a few days before that. There are a number of implications from those incidents that I think we should be more closely considering.
The very act of writing my usual shtick here at PJ Media was a problem due to that outage. Not that PJ Media was directly affected, but a goodly chunk of the rest of the internet certainly was, including my own website(s). ChatGPT and X were out, as well. Researching my story that day ended up being a challenge.
One of the advantages of the internet, when it first came along — you know, when Grok was still banging out emails with a rock and a slate tablet — was that it was decentralized. As designed, the system could deal with one or two players going dark for whatever reason, and still remain useful.
Thing is, like the American government, which was also designed as a decentralized entity, there’s always some putz who believes that things would go more smoothly if we were more centralized, allowing the economies of scale to take over. The very nature of that style of operation is that when problems occur, they’re huge ones.
Consider: One of the reasons plane, train, or bus crashes are such hot news items when they happen, and car crashes are not, is the sheer number of dead and injured. Yet, we’re told that “Mass transit is the answer.” (I note with interest that this is the claim made by large, centralized portions of the government, such as the federal DOT and the EPA.)
The result of this kind of thinking, of course, is the huge impact the outage of 11/18 had.
Another example: In my lifetime, there have been a handful of power outages involving millions. One memorable incident occurred on Nov. 9, 1965. Most of you weren’t alive then, but I have specific memories of the incident. The failure cut power to some 80,000 square miles and affected some 30 million people in parts of the Canadian province of Ontario, Connecticut, Delaware, Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, and Vermont. Tuning across the AM dial that night was an interesting experience.
In that case, the whole thing was caused by the failure of a protection relay in a hydro station in Queenston, some 70 miles west of me here, near Niagara Falls. I remember it was a cold and very clear moonlit night, and power demands were unusually high for the time. The family was just getting ready to sit down to dinner when the blackout hit. The relay failed because it was misprogrammed, and it was tripped by a small variation in the power coming from the Robert Moses power plant in the Falls. When the line it controlled went out, the load was redistributed to other lines, which, in turn, failed because of the spike in loading. The domino effect took over from there. It took 12 very cold hours to get it all sorted out.
Given the nature of the electric power network of the time, that kind of failure was pretty much baked into the cake. It shows the frailties involved with a centralized system.
I think it's safe to assume at this point that there was no malice involved in the Cloudflare outage, and that, much as with the 1965 incident, the cause was a series of failures. I really have no understanding of the situation beyond that. I suspect it was a maintenance cycle gone awry, but I have little to go with beyond that.
But what, I wonder, if someone actually wanted to bring the net to its knees on purpose? Perhaps some terrorist group, or someone wanting to make a political point. Or perhaps in preparation for an attack. In each of those cases, it appears to me that it wouldn’t take much to make something like that happen. What if, for another example, some government operative decided to shut down the net, perhaps because it is now, for the most part, our chief source for news?
The fact is that these days, just about everyone has their services at least incidentally run through Amazon Web Services or Cloudflare. It should be said that the folks at Cloudflare are brilliant, having found a way to place themselves at the very center of the ‘net and thereby a goodly percentage of everyone’s online business. Make lots of cash that way, given the volumes of traffic.
You can certainly make the argument that, in terms of reliability, you’re better off going that route than running your own data center and paying for the techs to run it and the software and hardware, etc. And obviously, it’s cheaper, as well. I would add that we simply don’t have enough qualified IT folks to run all those separate, smaller shops. Running a centralized operation sidesteps that issue, too. But with those advantages comes the problem that, in that configuration, it is much easier for someone to pull the plug on you and everyone else with you.
But in chasing that reliability, we have left ourselves open to the kind of massive failures we saw these last couple of weeks. More worrisome is the idea that it wouldn’t take much to bring the whole thing down. Get a few knuckledraggers to fly planes into a data center or two, and half of our national business model falls over, and our information streams with them.
If that’s a huge issue, if someone with very little tech experience can cause such havoc, imagine what a handful of highly trained IT guys — hackers, if you will, either in the employ of some government (including our own) or acting independently, as so many wishing for the demise of our country — can do.
It's time to give some thought to these things. We're up against some serious technological challenges from places like China, and it concerns me that so much of our hardware and software is dependent on China for it's origins and support. China knows our systems better than we do, in many cases. If it wanted to, it could cause outages that make the Cloudflare screwup the other day look like small potatoes. Eventually, they're going to figure that out.
Join PJ Media VIP for deep analysis, investigative originals, and commentary that refuses to bow to the cultural trends reshaping American life. Use the promo code FIGHT to get 60% off your new one-year membership!
Join the conversation as a VIP Member