Service · disaster recovery

Disaster recovery colocation in Amsterdam: a real second site

DR is a distance problem before it is a shopping problem. Far enough from your primary to survive the same incident, close enough that replication still works, on a different power feed and a different fibre path.

What you get

A second site you can fail over to.

A second site at a sane distance

Far enough for a different power feed and a different flood zone, close enough for synchronous writes. We put numbers on both ends.

Distance

A fibre path you can name

Dark fibre or a wavelength between campuses, not a best-effort tunnel over the public internet. Diverse ducts where the operator can prove them.

Dark fibre

Sized for failover, not for show

One rack of cold standby or a full mirror. We size it from your RTO and RPO, not from your primary’s rack count.

RTO / RPO

A runbook someone has tested

Who declares, who fails over, who tells the auditor. We help write it and sit in the first test; you own it.

Runbook

Specs that matter

Distance, latency, and what they cost you.

The two sites are one design. Pick the distance first and the replication mode follows; pick the replication mode first and you may be shopping for a building that does not exist.

Distance between sites: Roughly 10–30 km across the Amsterdam metro: different substation, different flood zone, same time zone as your staff.
Synchronous replication: Comfortable under about 1 ms round-trip, so within the metro. Every added millisecond lands on your commit latency.
Asynchronous replication: Anything further. Your RPO becomes the replication lag, so measure the lag rather than trusting the brochure.
Inter-site fibre: Dark fibre or a wavelength between campuses, ordered through the operators. Ask for the duct route, not just the endpoints.
Footprint: From one rack of standby gear to a mirror of production. Most start small and grow after the first real test.

How we place it

Four steps, from RTO to first failover test.

01
Tell us what you needWhere your primary sits, your RTO and RPO, and what genuinely has to run after a failover.
02
We engineer and negotiateWe pick a second campus on power feed, flood zone and fibre path, then price the link with the space.
03
We fast-track the introductionBoth operators get the same brief, so the cross-campus fibre is quoted alongside the racks instead of two months later.
04
You contract with the datacenterTwo contracts, two SLAs, one runbook. We stay on for the first failover test.

Read the full process

Who we’ll send elsewhere

Not a backup product, and not a second rack next door.

We do not sell backup software, replication licences or managed DR; we place the site and the fibre. If what you need is off-site backup and a weekend of restore testing, we will say so and you will save a lot of money. And a second rack in the same hall is not disaster recovery: it survives a server, not a substation. If you need true geographic separation from Amsterdam, another metro is the honest answer and we will tell you that too.

Questions we get

Distance, replication mode and who fails over.

How far apart should the two sites be?

Far enough for a different power feed, fibre path and flood zone; close enough that synchronous replication still fits inside your write latency. Across the Amsterdam metro that is roughly 10 to 30 km.

Synchronous or asynchronous?

Synchronous if your database cannot lose a committed write and the round-trip stays near a millisecond. Asynchronous everywhere else, with your RPO set to the lag you can live with.

Can my primary stay in my own server room?

Yes, and plenty of placements start that way: keep the server room, put the recovery site in a datacenter, then move the primary later once the move is boring.

Who runs the failover?

You do. We help write the runbook, sit in the first test, and stay reachable as technical contact afterwards.