June 3, 2026
Name shame in the server room
DNS Is for People – Not for IT Infrastructure
Engineers say ditching internet phonebooks for office systems could cause total chaos
TLDR: The article argues companies should stop relying so much on DNS, the system that turns names into internet addresses, for internal tools because it can become a single giant failure point. Commenters mostly pushed back hard, saying hard-coded addresses and hosts files sound messier, riskier, and way more painful to maintain.
A spicy tech essay tried to make one big argument: the internet’s address book should be for humans, not the behind-the-scenes machinery that keeps companies running. The author says names are great for public websites, but inside a company, every extra lookup system adds another way for things to break. The warning taps into an old IT meme — “It’s not DNS… it was DNS” — and points to famous meltdowns like Meta’s outage, where one failure spiraled so badly people reportedly struggled with building access. The pitch? Skip the name system for internal tools, use direct addresses, or even old-school hosts files instead.
The comments, however, came in like a mob at a reality-show reunion. One camp treated the whole post like a full-blown anti-DNS crusade, with critics saying replacing one central system with a pile of hand-managed machine lists is not simplification — it’s administrative fan fiction. The loudest gasp came over the suggestion to “just use /etc/hosts,” which one commenter mocked as basically creating “bespoke DNS servers” on every computer. Others rolled their eyes and said, sure, hard-coded addresses sound cute until you need to move a service and suddenly every machine needs updating.
Still, the thread wasn’t all outrage. There was classic nerd humor, some delightfully niche trolling about using even more DNS, and the usual haunted refrain that no matter what goes wrong online, somehow the address book drama is never truly offstage.
Key Points
- •The article says DNS is well-suited for public-facing services because it gives humans memorable names and allows IP changes without affecting clients.
- •The author argues that using DNS for internal IT infrastructure adds a dependency that can increase failure risk and outage blast radius.
- •The article cites high-profile incidents, including the Facebook/Meta outage, as examples where DNS-related dependencies magnified operational impact.
- •It identifies DNS caching and TTL behavior as an operational challenge because clients may not pick up changed IP addresses quickly or consistently.
- •The article proposes direct IP configuration and provisioned /etc/hosts entries, managed with tools like Ansible or pyinfra, as alternatives for internal systems.