Service · Backup and restore

Cloud backup

Fire, flood and burglary draw no distinction between a server and the backup appliance standing beside it. Closing that gap is what a remote copy is for, and it also retires the weekly ceremony of somebody slipping a disk into a bag on the way out, or forgetting to. We place that third copy with a host inside the European Union, and the encryption happens on your side of the firewall.

Third copy
the off-site leg of 3-2-1
Encrypted
on your side of the firewall
Object lock
deletion refused until expiry
Europe
the region is named in writing

Where this service reaches

Treat it as a partner to the NAS. Everyday mishaps are repaired from local disks within minutes; the remote copy exists for the day nothing local is left.

Put your scope to an engineer

Host and region

S3-compatible capacity from OVHcloud, Scaleway or a comparable European operator, object locking switched on, and a processing agreement your data protection officer reads before anyone signs. The region appears in the documentation.

Keys in your hands

Encryption runs before a single byte leaves the building. You hold the key; a spare lives in a safe or in the board's password vault, and the documentation states exactly where.

Traffic shaping

Upstream bandwidth gets measured first, since business fibre is usually far weaker in that direction, then the job is capped so that neither video calls nor order entry suffer during trading hours.

Retention and the bill

Ageing copies drop into cheaper classes, then vanish on expiry. We also price a full recovery, a figure that comparison tables omit with remarkable consistency.

A recovery scenario

Which system wakes first, how long each is expected to take, and how machines are stood up at the host when the originals no longer exist. A few pages, written to be read by somebody under strain.

Something to show an auditor

Job history, dated recovery notes and a table of measured durations. Enough for an insurer, for a review against ISO 27001 principles, or for the continuity questions attached to NIS2.

The way an engagement runs

Days rather than weeks to configure. The opening upload is another matter, since a few terabytes over ordinary business fibre will happily occupy more than a week.

01

Arithmetic

Volume, retention and available bandwidth. Those three produce a monthly figure and an upload duration, both known before you commit.

02

Plumbing

Capacity created, keys issued with the narrowest rights available, encryption enabled, object lock armed, lifecycle rules written down.

03

Seeding

The opening copy runs at night and across weekends, or under a speed cap, leaving the line usable by human beings.

04

Proof

A sample comes back down onto an isolated machine. No successful recovery, no sign-off.

No cloud recovery will ever outrun your download speed. Shifting a terabyte across a 100 Mbit/s line occupies most of a day, and services still have to be started afterwards. Read that as a reason to keep both copies rather than to distrust the remote one. Minor accidents are repaired locally in minutes; the off-site copy exists for the scenario in which nothing local survives at all.

Questions and answers

Occupied storage is billed, and several hosts bill retrieval too. Volume, retention and the host's price list settle the sum, which is why we produce it before anything is switched on. Invoices grow out of hand in exactly one way, through copies nobody ever taught the system to delete, so expiry rules go in on day one.

Not with the hosts we pick, and the documentation names the region so you can check. Backups almost inevitably hold staff and customer records, and moving those to a third country brings further obligations under the GDPR. Staying inside the Union is simply the path of least trouble.

Not while the configuration holds, since the material is encrypted before it departs and the key stays in-house. That cuts both ways, naturally, because a mislaid key turns the archive into noise. Key custody is documented for precisely that reason and re-checked whenever a rehearsal runs.

Priority systems are rebuilt as machines at the European host while staff carry on from home across an encrypted link. The sequence, and the expected duration per application, are settled in advance in the recovery scenario, which we revise after every rehearsal.

Put one copy beyond your walls

Send over the volume needing protection and what your connection can push upstream. Back comes a monthly figure and a realistic seeding time.

When we are around
Weekdays, 8:00 to 18:00 CET; answers land inside one working day
Talking it through
A call on Teams or Google Meet, whenever writing is not enough

We set strictly necessary cookies only: they keep the site running and remember the city you chose. Nothing here is used for advertising or tracking. More in our privacy policy.