— Servers · MCO
Servers — they run, and they run well
A server isn't just a machine. It's uptime, performance, security and backup. We handle all four.
— Partenariats & certifications
Certifications systems & servers.
— Servers in numbers
Three figures to gauge our rigour.
3-2-1-1-0
Immutable backup strategy, tested quarterly. 3 copies, 2 media, 1 off-site, 1 immutable, 0 restore errors.
< 2 h
RTO validated in real DR exercises, not a paper promise. Documented, tested, up-to-date recovery plan.
0
Unplanned reboots over 12 consecutive months for clients under MCO. Cadenced patch management, 24/7 supervision.
BISPRO data, July 2026
01 · ARCHITECTURE
Architecture & deployment
We size right: no superfluous overlay, no under-capacity that wrecks ROI.
Bare-metal vs. virtualisation
Proxmox VE, VMware vSphere, Hyper-V depending on context and budget.
Right-sizing
CPU, RAM, storage, IOPS, no vendor-driven oversizing.
Storage
NAS / SAN, suitable RAID, snapshots, deduplication, immutability.
Identity & directory
AD / Entra ID, hybrid, GPO, Kerberos hardening.
Linux servers
Debian, Ubuntu LTS, RHEL / Rocky in production.
02 · VIRTUALISATION
Virtualisation and consolidation
One physical server per application means capital tied up and failures multiplied. Virtualisation groups your services onto a controlled hardware base, with backups and restarts that finally become simple.
Proxmox VE — no licence, no ceiling
An open-source hypervisor we have deployed and run for years. No per-socket or per-core licence: the budget goes into hardware and backup, not into an annual subscription that keeps rising. Built-in backup, replication between nodes, live migration of virtual machines.
Hyper-V and VMware where context demands
Some business software vendors only support one specific hypervisor, and some IT departments have established standards. We work with all three and tell you plainly what each really costs over five years, licences included.
Consolidating what exists
Moving ageing physical servers onto virtual machines, without reinstalling the applications where that is possible. Less hardware to maintain, less power drawn, fewer parts to fail, and restores that take minutes instead of a day.
Clustering and high availability
Two or three nodes with shared storage mean a hardware failure cuts nothing: virtual machines restart automatically on another node. We size this against what an hour of downtime actually costs you, not against a catalogue.
Virtual desktops
Where business applications justify it, desktops can run on the server and display on thin clients. Data never leaves the server room, a stolen machine contains nothing, and replacing a workstation takes ten minutes.
03 · BACKUP & DR
Backup & disaster recovery
Backups without tests are fiction. Every contract includes restore drills.
3-2-1 & 3-2-1-1-0 policy
3 copies, 2 media, 1 off-site, 1 immutable copy, 0 errors after verification.
Endpoint + server backup
RPO matched to criticality, contractual RTO, traceable logging.
Restore tests
Quarterly, documented, played on simulated incidents.
Formalised DR plan
RTO / RPO, runbook, contacts, escalation. Updated annually.
04 · DIRECTORY & MAIL
Directory, mail and file sharing
The building blocks that shape daily work: who is allowed to do what, where the files are, how mail leaves. Set up badly, they produce years of workarounds and permissions open to everyone.
Active Directory and permissions
A clean directory, coherent groups, rights granted by role rather than person by person. A leaver is handled in one operation, so is a joiner. Without this, permissions accumulate on accounts over the years until nobody knows who can reach what.
Business email
Microsoft 365, mail hosted in France, or an in-house server depending on your constraints. SPF, DKIM and DMARC records configured so your messages land in the inbox and your domain name cannot be spoofed.
File sharing and NAS
Structured workspaces, permissions by department, version history to recover a file overwritten by mistake. NAS devices are backed up like everything else: a NAS is not a backup, it is one more server to protect.
Remote access to business applications
Secure publication of applications for home working and remote sites, over VPN with two-factor authentication. Never a remote desktop opened directly onto the internet: it is the most exploited way in during attacks on smaller businesses.
05 · MCO
24/7 supervision & MCO
Monitoring is only useful if it triggers action. Filtered alerting, escalation, response.
CheckMk + BISPRO Monitor
Centralised multi-site supervision, lightweight agents, OS / hypervisor / application metrics.
Alerting
Multi-channel, SLA escalation, noise suppression.
Patch management
Monthly or weekly depending on service criticality.
Monthly report
Availability, incidents, priced recommendations.
06 · BY CONTEXT
What changes with your line of work
Sizing, tolerance for downtime and regulatory obligations vary sharply between sectors. This is how we adapt the infrastructure.
Industry and manufacturing
Servers driving shop-floor supervision or production planning do not tolerate unplanned restarts. Isolation of the production network, protected power, spare parts on site and maintenance windows aligned with line stoppages.
Healthcare and social care
Health data subject to strict retention and traceability rules. In-house hosting or an approved provider, access logging, encrypted backups and tested restores to satisfy accreditation audits.
Local authorities and public bodies
Public procurement, retention of official records, continuity of service to residents. We work with sovereign infrastructure and document what an inspection requires, including retention periods and access rights.
Practices and regulated professions
Professional secrecy, retention obligations, a document base that keeps growing. Encryption, strict separation of client files and off-site backup so that neither a fire nor ransomware takes the archive with it.
Multi-site companies
Servers centralised or distributed according to link quality, replication between sites, and a recovery plan that allows for losing a site entirely. One point of contact for the whole estate.
07 · FAQ
Frequently asked questions
Going further
Audit my server infrastructure.
Inventory, performance, backup, exposure. Actionable report within 5 business days.
Financing up to 63 months available* — *subject to approval

