Boring by default
New technology earns its place. Proven technology keeps the lights on.
About Atlantika
Atlantika has been running other people's critical systems since 2017. We are deliberately small, deliberately boring, and unusually easy to reach.
Our story
In 2017, three engineers who had spent years patching around other people's outages rented a single rack in East London. The plan was modest: build the infrastructure we wished we could buy — clear pricing, honest maintenance windows, and a phone number that a human answers.
Nine years later we operate three core data centres and nine edge locations, but the shape of the company has not changed much. We still run the network ourselves. We still publish incidents. And we still think the best infrastructure is the kind you forget about.
Today the team is 23 people across the UK, Netherlands and Germany. We host retail platforms, healthcare back ends, media archives and a surprising number of logistics systems that keep moving at 03:00.
“We moved our order platform to Atlantika in 2024 and have not had a 3 a.m. call since. That is the whole review.”
Principles
They are not framed on a wall. They show up in runbooks, change reviews and the way we answer the phone.
New technology earns its place. Proven technology keeps the lights on.
We do not blame the carrier, the hypervisor or the firmware. We fix it and write it down.
Runbooks, decisions and incident reviews live in the same place. Memory is not documentation.
Support is staffed by engineers. There is no tier-one script between you and someone who can fix it.
A short history
We have grown by adding racks and people, not by buying other companies or renaming the same service three times.
Two customers, one very second-hand KVM, and a strict rule about writing things down.
Layer-3 connectivity across both London halls, with customer-managed ACLs from day one.
Our first 10G cross-connects and a proper European peering footprint.
Annual penetration testing and a formal access-review process, without the certificate theatre.
Manchester and Dublin came online, mostly because customers kept asking for them.
Built on the same hardware estate, priced per gigabyte rather than per question.
99.98% trailing availability and dual-stack across twelve locations.
How we work
We do not grow faster than we can train people. Every engineer spends their first month on the support rota, and every customer-facing promise is checked against the on-call calendar before it is made.
Want the detail? Our infrastructure page covers the physical side, and the security page explains how we handle access and disclosure.
Leadership
A flat team with clear ownership. If something breaks, you will know who is working on it.
Managing Director
Runs the commercial side and still reads every incident review. Joined in 2018 from a network operator.
Head of Infrastructure
Designed the current network and has strong opinions about cable labelling. One of the three founders.
Principal Network Engineer
Peering, routing and the person to call when BGP misbehaves. Previously at a London transit provider.
Head of Customer Operations
Owns support, onboarding and the on-call rota. Believes a good handover is a form of kindness.
We are happy to talk through our network, our on-call process or a migration you are dreading. No slide deck required.