A VPS is a slice of a data centre machine reserved for you alone: so many cores, so much memory, a disk of a stated size. Three families of workload tend to land there. A finance or stock package that staff open over a published desktop. A site or storefront grown too heavy for its shared plan. And one small in-house tool, knocked together in haste years ago, that nobody can now work without. Billing is monthly and follows the capacity actually consumed, and moving up or down a size relocates nothing.
Renting the machine takes minutes and ends up as a monthly invoice. The real work happens inside it: the operating system, access locked down, your applications, the copies, and somebody reading the graphs before the disk fills up.
Cores, memory and disk are derived from load measured on your current setup, month-end peak included. We happily start one size down, because adding capacity takes ten minutes whereas an oversized plan is paid for every month for three years.
OVHcloud, Scaleway, Outscale or Clever Cloud, weighed on price, the region you need, any certification required and whatever your business already uses. The account is opened in your company name, never ours.
Either Linux or Windows Server, logging in by key or behind a second factor, the host firewall switched on, and management ports shut long before anyone signs in.
Sage, Cegid, EBP or Divalto reached through a gateway with strong authentication, and certainly not via port 3389 sitting open to the whole internet.
A snapshot before every update so a bad change can be undone in minutes, plus an independent copy with a different European provider that we actually restore from.
Uptime, load, free space, expiring TLS certificates. Security fixes follow an agreed calendar rather than arriving on the day something breaks.
Lead time depends on the provider and on how much has to be carried over. Data copying is scheduled for the evening so nobody stops working.
What has to run on the machine, how many people will be connected at once, which licences you already hold and what must be brought over from the present setup.
We raise the instance inside the account belonging to your business, install the system, close off access nobody needs and fit the monitoring agents.
Applications are installed, databases and files move across in an agreed window, then workstations are configured to connect.
Credentials are handed over through a password manager, along with a one-page description of the environment and a restore performed live in front of you.
An exposed remote desktop port does not survive a single day. Bots sweep provider address ranges without pause, and a fresh machine with 3389 open receives its first password attempts within hours. A great many ransomware cases begin exactly there, as CERT-FR keeps pointing out in its bulletins. So published desktops never face the internet head-on: they sit behind a VPN, or behind a gateway demanding a second factor, and nothing else.
It follows from what the machine has to do. A site, a shop or an interface for other systems runs more lightly on Linux and costs nothing in licences. A finance package built for Windows, or remote desktops for the team, calls for Windows Server together with its access licences, usually rented monthly from the provider. We cost both before deciding.
Two separate lines. The provider subscription varies with the size of the machine and the licences on it, and they bill you for it directly; we take no margin, and published rates change too often to be copied onto a web page. Our line is the administration, at EUR 95/hour excl. VAT for time spent, or folded into a Start, Business or Premium plan. Both figures are given to you before you decide.
Yes, and it ranks among the commonest reasons for moving at all. After signing in over the VPN or gateway, they see exactly the screen they had at the desk. The remote working agreement or charter is a matter for your HR people; the technical half is ours, so that it is not what holds things up.
First we work out why. A badly indexed database, or a scheduled job recalculating everything every minute, will eat whatever headroom exists, and moving up a size would simply postpone the problem at a higher price. Where capacity genuinely is short, we step up a size, divide the roles between two machines, or recommend dedicated hardware.
Your company does. The contract, the billing and the domain name stay in your name and we work through a separate administrative login. If you ever decide to hand the running of it to somebody else, there is nothing to buy back and nothing to move: you simply withdraw our rights.
Tell us what has to run on the machine and how many people will connect to it. We come back with the resources needed, both cost lines and a realistic launch date.
Message received
An answer follows inside one working day. Report an outage that is stopping people working and it moves ahead of everything else.
Nothing here under that name. Check the spelling, or simply choose the nearest large city instead. Since every engagement runs remotely, whichever you pick changes nothing about what we do for you.
