Another One Bites the DNS
A week ago AWS suffered a major outage caused by a DNS error, leaving numerous services and apps inaccessible for hours on a global scale. In my report on it, I stated that corporate monoculture is going to become a greater issue, since only a handful of tech giants wield the vast majority of the power in how the internet’s functionality is handled. Yesterday, I was proved correct.
Around noon EST, Azure Front Door, Microsoft’s cloud content delivery network, went down. And stayed down until the configuration change that caused the outage was reverted, some 8 hours later. (Which strikes me as technospeak for ‘restored to a previous version’, a rather concerning happenstance in and of itself, that a flaw can be so catastrophic that there’s nothing else to do but undo an entire update.) Those downtime hours were a whole business day, and the disruption of services was as far reaching as healthcare and airlines, XBOX games, as well as nearly all of Microsoft products. I myself could not access Minecraft on PC and that’s how I discovered there was an issue in the first place. My first stop upon being unable to login was to check the downdetector. Fully half of the services listed on the main page were affected. Some outage spikes were brief, but most were still ongoing at the time I checked (a little after 2 pm).
We talk about client trust and how it can be lost. Reddit’s having a field day with this, following so closely to the AWS outage. There is a tone of exasperated ridicule in the reaction. A sidelong smirk and nudging elbow of ‘of course it’s a DNS error’. A sense of inevitability because all the eggs are in one basket that is being increasingly carried by AI. Comments range from jokes referencing ‘The IT Crowd’ to accidentally unplugging the internet because they tripped over it or had to vacuum.
I titled my report on the AWS outage ‘Glass Cannons’ for a reason. Sure, these tech giants are powerful and have integrated themselves into products and services that have nothing to do with their base operations, allowing them to profit hand over fist. But they are fragile. A DNS error should not take down half the internet, for crying out loud. This is a stark example of why diversification of the infrastructure is essential. One comment I read this morning posed the question to the effect of imagine that this was a malicious attack, and what would the fallout be? That’s terrifying.
They say disasters come in threes, so who will be next? Who’s left to be next? And what will it take to break through the idea that infinite growth is sustainable when it comes at the expense of fundamental security and infrastructure? I’m just an analyst, I don’t have the means to create a solution to the problem. But I can recognize the pattern. Cloud computing is a great way to do business, but it needs a literal spreading of the wealth. After all, the bigger they are, the harder they fall.
Posted on LinkedIn, 10/30/25
