Service · Servers in NIT data centres

Server installation and management in the main data centres del Northern Italy.

We support Italian and foreign companies in operations within Northern Italy datacenters: from the installation of AI servers with NVIDIA H100/H200 GPUs to 24/7 maintenance, from remote-hands physical access to the implementation of security systems. Vendor-independent, we operate within your colocation facility as an extension of your IT team.

02 / AI & GPU server installation

AI & GPU server installation

  • Rack-mount hardware: NVIDIA H100/H200 SXM/PCIe, AMD MI300X, A100 80GB; servers Dell PowerEdge XE9680, HPE Cray, Supermicro AS-8125GS, Lenovo ThinkSystem SR685a V3
  • Power & cooling profiling: verification of kW budget per rack (30–50 kW high-density), support for Direct Liquid Cooling and rear-door heat exchanger
  • High-bandwidth networking: cabling 10/25/100/400 Gbps, InfiniBand HDR/NDR, RoCE v2 for multi-node training, LACP bonding
  • Firmware bring-up: AI workload BIOS tuning, BMC/iLO/iDRAC, secure boot, IPMI hardening, baseline firmware verification
  • OS bare-metal & runtime: Ubuntu 22.04/24.04 LTS, RHEL 9, Rocky Linux; NVIDIA driver + CUDA toolkit, NCCL, MIG, container toolkit
  • Orchestration: Kubernetes with GPU operator, Slurm for HPC, Run:AI, Determined.AI for training queue management
  • Validation & benchmark: stress-ng, gpu-burn, NCCL all-reduce, MLPerf Inference baseline, P2P bandwidth test
03 / Remote-hands & data centre access

Physical access via remote-hands

  • Access procedures: badge / biometrics / escort on your behalf, visit-pass management with the data centre NOC
  • Response SLA: within 2h for the Milan area, 3–4h for extended areas; on-call 24/7 for emergencies
  • FRU replacement: drives, PSU, RAM, GPU, network cards; module reseating, PCIe seating verification
  • Cabling & labelling: tidy patch panels, TIA-606 labelling, up-to-date mapping, OM4/OS2 fibre dressing
  • Console & power cycling: serial console, KVM-over-IP, controlled power cycling with PDU verification
  • Photographic audit: pre/post intervention report with photos and detailed ticket, client hand-over
04 / Systems support 24/7

Operations and ongoing support

  • Monitoring stack: Zabbix, Prometheus, Grafana, Netdata; SNMP, IPMI, custom exporter integration
  • Multi-channel alerting: PagerDuty, Opsgenie, email/SMS/Telegram, multi-level escalation policy
  • 24/7 SLA: response time ≤15 min, restoration ≤4h for critical incidents (Sev-1)
  • Patching & hardening: agreed maintenance window, rollback plan, CIS Benchmark, STIG on request
  • Log management: ELK Stack, Grafana Loki, Splunk; compliance retention, queryable from client console
  • Reporting: monthly report: uptime, MTTR, MTBF, RCA for major outages, capacity trend
05 / Physical & cyber security

Security systems implementation

  • OS hardening: CIS Benchmark Level 1/2, STIG, ANSSI; audit-ready configuration
  • Identity & access: MFA on all administrative access, jump host, ZTNA (Zscaler, Cloudflare Access, Tailscale Enterprise)
  • Network segmentation: Zero Trust microsegmentation, identity-based firewall, VRF for multi-tenancy
  • Immutable backup: Veeam Hardened Repository, WORM object storage (Wasabi, Backblaze B2, Azure Blob immutable), air-gap
  • EDR / XDR: CrowdStrike Falcon, SentinelOne Singularity, Microsoft Defender XDR, Sophos Intercept X
  • Vulnerability management: Nessus, Qualys, Rapid7 InsightVM; recurring scanning + remediation tracking
  • Compliance: preparation for your ISO 27001 audit, NIS2, GDPR; operational documentation ready for auditors
06 / Datacenter Northern Italy coperti

Where we operate on-site

We work in the main datacenters of Northern Italy on behalf of clients who have colocated their infrastructure there. We operate as an independent third party: no affiliation with the datacenter operators, no lock-in.

Milan metropolitan area

Equinix ML1–ML6 Caldera · Milan Global colo
Data4 MIL01–MIL03 Cornaredo (MI)
Stack EMEA MIL01–MIL04 Settimo Milanese (MI)
Digital Realty MIL1–MIL2 Caldera (MI) formerly Interxion
Vantage Data Centers MIL01 Milan
IRIDEOS Avalon Caldera (MI)
CDLan Caldera Park Settimo Milanese (MI)
WIIT Settimo Milanese (MI)

Province of Bergamo

Aruba IT3 Ponte San Pietro (BG) Tier IV

Province of Pavia

Supernap Italy SUP1 Siziano (PV)

Province of Brescia

Open Hub Med Brescia

Valle d'Aosta

Engineering D HUB Pont-Saint-Martin (AO)

Distributed offices

TIM Sparkle Milan · Turin Metropolitan offices

Have you colocated in a data centre not on our list? Contact us anyway: we operate outside the listed facilities subject to access procedure verification.

07 / How operiamo

Intervention methodology

  • Pre-intervention audit — target hardware inventory, scope-of-work, window scheduling, ticket opening with the data centre NOC
  • Coordination — alignment with client IT manager + data centre operations, confirmation of access and PDU
  • Execution — on-site team with real-time photographic report via ticket, continuous communication with the remote client
  • Validation — smoke test, performance baseline, verification of active monitoring, cabling confirmation
  • Hand-over — updated LLD, runbook, escalation contact list, final photographs; client sign-off
  • Maintenance — entry into the monthly or on-demand on-call rota, continuous monitoring if subscribed
08 / FAQ data centre & remote-hands

Most frequently asked technical questions.

Quali datacenter del Northern Italy coprite per remote-hands?+
We operate in the main facilities: Aruba (BG), Equinix Milan (ML1-ML6), Data4 Cornaredo, Stack Settimo Milanese, Digital Realty, Vantage, Supernap (PV), IRIDEOS Avalon, CDLan, WIIT, Engineering D HUB (AO), Open Hub Med Brescia, TIM Sparkle. Also unlisted data centres, subject to access procedure verification.
AI GPU server installation: which stacks do you support?+
Bare-metal Ubuntu/RHEL/Rocky with NVIDIA drivers + CUDA + NCCL + container toolkit; orchestration with Kubernetes (GPU Operator), Slurm, Run:AI, Determined.AI. Hardware: Dell XE9680, HPE Cray, Supermicro 8125GS, Lenovo SR685a with H100/H200 SXM, A100, AMD MI300X GPUs.
Average remote-hands time for the Milan area?+
≤ 2 hours from request for data centres in the Milan metropolitan area (Caldera, Cornaredo, Settimo Milanese, Ponte San Pietro). On-call 24/7 with 15-minute response SLA for emergencies.
Do you also work for foreign clients with servers colocated in Italy?+
Yes. Many European customers (UK, DE, FR, CH, AT) have colocation in Italy and use us as local remote-hands. Communication in English, tickets via Jira/ServiceNow/email, standard EU B2B invoicing.
Vendor independence: what does it mean in practice?+
We do not sell licences directly, we have no exclusive reseller agreements, and we receive no commissions from manufacturers. We recommend the architecture best suited to your use case even if it goes against the market "default".
24/7 SLA for ongoing systems support: how does it work?+
Monthly or annual contract, ≤ 15 min response time 24/7 for Sev-1 incidents, restoration ≤ 4h. On-call rota across the senior team. Monthly uptime/MTTR/MTBF report + documented RCA for every major outage.
Security implementation: do you work only on the cyber side or also on the physical side?+
Both: on the physical side we assist with badge access, biometrics, CCTV, rack segregation; on the cyber side with OS hardening (CIS/STIG), MFA/ZTNA, microsegmentation, EDR, immutable backup and ISO 27001/NIS2 audit preparation.
Let's talk

Ready to talk with a systems engineer?

You will always be answered by a technician, never a call centre.

02 95 17 550