A technical buyer evaluating hosting does not want slogans. He wants hardware facts and an SLA. If those are missing, the usual assumption is that something is being hidden. We prefer the opposite: put the numbers on the table.
Hosting662 runs in our own low-density 662 datacenter in the Czech Republic. Cooling is designed as 1+1 with 100% reserve capacity and a tight temperature band of 20/23 °C. Every outlet on both power circuits is UPS-backed; if mains power fails, an outdoor gas generator starts immediately — and that backup path also covers cooling. For the facility as a whole (internet, network, power, air conditioning) we publish an SLA of 99.99% with a 20-minute response time, classified as TIER II for disaster scenarios.
On the VPS side, each virtual server lives in a cluster of four physical dual-CPU hosts. Storage is not a single disk under one machine: it is CEPH across the whole cluster. We operate more than three such clusters. That architecture is what makes live migration without downtime possible during maintenance — the VPS can move while customers keep working. Internet connectivity to the service is 2× 10 Gbps, again with 99.99% SLA and a 20-minute response target.
The physical layer is intentional too: HPE DL360 servers, SSD and U.2 media feeding the CEPH pool, and 10 Gbps networking. Backups run daily into a separate encrypted backup store; on a support ticket we can restore up to a week back. Customers who want an extra safety net before a change can add self-service snapshots and restore them anytime from the client panel.
Put together, the stack is built for two outcomes that matter in production: unnoticeable migration during maintenance, and fast recovery if something goes wrong with customer data. And at this VPS price tier, CEPH is the differentiator most competitors simply do not offer.
Full parameter list: Technical specifications. Practical answers for customers: FAQ on the VPS hosting page.