Why We Should Cheer the Facebook Crash
Posted October 06, 2021
I’m happy to report: I wasn’t even slightly nonplussed by the Facebook crash. I hope you were not either.
I hardly use it. And yet it is like cigarettes: you know they are bad for you, but it’s a hard habit to shake. So long as people check their Facebook accounts at least a few times a week, the company has a business model and can brag to advertisers. The revenue model depends on people not leaving. They don’t have to love it; they just can’t leave it.
So I admit to having a few accounts, not to learn anything but mostly to push content. It’s become ever less useful for that purpose but it is not yet useless. Millions, even billions, are in my position.
Instagram I can take or leave. I do like WhatsApp but there are plenty of other means of communicating, such as the privacy-oriented Telegram and Signal. And there’s good ‘ol SMS which always seems to work. Plus most laptop workers today communicate via Slack or an impersonating platform.
The rest of the world seemed to freak out about the crash of this suite of services for 6 hours. I get it. People take them all granted as if they work automatically and nothing can ever go wrong. That’s completely untrue. It’s good for the world to be reminded that this is just another company stitching together wires and code and it can screw up as much as you or I can.
I’ve long experienced schadenfreude when large companies have websites that crash. That’s because I’ve been responsible for uptime at several small and medium-sized companies. Web work is hard. Servers go down. Everyone freaks out and points fingers. I’ve had lines of people outside my office, angry employees demanding their instant fix now. It doesn’t help to get it fixed to get mad, but that’s the way it always goes.
So when Google goes down or Citibank crashes I like to look at them and say: see, it happens to everyone! Doesn’t make it any better.
If I set up a server in my garage and link up the neighborhood to it, and it crashes, I certainly bear responsibility for it. I’m the one point of contact. If my machine dies, everything dies. In network parlance, this means there is a single point of failure. One thing goes wrong and the whole system is done.
In the old days, this was the only way things could work. But starting in the first decade of this century, there was a new innovation called distributed software. I run an instance of it and you do too. We can connect directly with each other. If you crash, I can still communicate with others on the network. If we are running the same protocol, we have a distributed system that is less prone to takedowns. If there are thousands and millions of them, it becomes a multi-headed hydra that can’t go down.
That is certainly not Facebook. In fact, I’m not sure we knew until the last day or so just how preposterously centralized this company truly is. When the main server got confused about routing issues, everything went down. Not just Facebook, but every company that Facebook owns was hit.
Gosh, I would have loved to have been a fly on the wall at Facebook headquarters during those six hours. When a company’s entire world crashes, the reason is never obvious. There is no software out there that screams the answer. You look at symptoms, check traffic, check logs, use diagnostic tools, and you try to do all of this in a calm and systematic way, so that you are not changing things at the same time. It’s a process of elimination that is rooted in logic and reason. Freaking out accomplishes absolutely nothing.
Initially, the company faced a complicating factor. All its communications systems run on Facebook’s servers. They were down. So people could not even speak to each other. They resorted to phone calls, frantic ones. They used other tools including email. Most of their employees are still remote so it was enough of a task just to find out who possibly could figure this one out.
It must have been total chaos — the kind of chaos that results from a failure from a single center point.
Yes, the company lost billions, but without any greater effect on markets generally. Mostly it was just an annoyance. What the company has lost is the credibility that comes with being on the cutting edge of innovation. With this kind of structure, there is simply no chance for innovation. Nor will there be a way for management alone to rectify this. The structure of centralization is too embedded in their operations. There are too many switches and wires to roll that back now. Plus too much disruption from within risks a repeat of the crashes, and that can never be risked again.
Thus did the world of tech remind us of why the blockchain matters. The world largely fell apart by force in 2020. Everything was challenged. But you know what did not even pay attention at all? Bitcoin. The entire cryptosphere. This is because it lives on a distributed system. It’s hardly the only new technology to enjoy this luxury, but it is the most famous one.
It cannot be taken down, not by governments and not even by silly human errors, such as one person uploading a bad update to a server.
The structure of information systems is undergoing an upheaval. George Gilder has already explained it all, but, as always, he is generally five to ten years ahead of the curve. Bottom line is that Facebook is destined for the dustbin of history, but the slide to the point will be slow, and its demise will gradually overlap with its many replacements in ways that most people will barely perceive.
Facebook will recover from its disaster, except that this incident is a sign to which the whole of the tech world paid attention. The model, the structure, the top-down management, it all worked in the past, but these days we know for certain what works much better.