Case study · Cloud Computing
A large Czech hosting provider with more than 50,000 customers moved from leased servers with no redundancy to enterprise-grade infrastructure. We built a 100 GbE MLAG network, a Proxmox cluster with 5,120 AMD EPYC cores and 280 TB of ZFS NVMe storage, migrated customer services and now run the whole environment.
The client is a large hosting provider operating in the Czech Republic. It serves more than 50,000 customers with shared hosting, VPS and dedicated server environments. Rapid customer growth exposed the limits of its existing setup: classic dedicated servers with no redundancy, no enterprise-grade security and no high availability. The client’s name is covered by a non-disclosure agreement.
The client ran on leased infrastructure that lacked:
For a hosting business this is a direct commercial risk: downtime, loss of customer data, no resilience to failures and difficulty meeting SLAs. Our task was to build modern, enterprise-grade server infrastructure from the ground up on proven technologies — Fortinet, Proxmox VE, ZFS, Proxmox Backup Server and Supermicro servers — then migrate customer services onto it and take over its operation.
We designed the project using our standard approach, in three layers: network → compute → storage.
The client now has high availability across its entire infrastructure, with redundant network, storage and compute, so its hosting services run reliably 24/7. 5,120 cores and more than 46 TB of RAM support thousands of VPS instances and customer environments, and the architecture lets the company add resources smoothly as it grows. ZFS combined with high availability gives customer data a high level of protection. Services for more than 50,000 customers now run on the new infrastructure, and Remote Admin handles its administration and maintenance, so the client’s team can focus on sales and customer care. Supermicro BigTwin servers and AMD EPYC Turin processors are among the most energy-efficient server platforms on the market, which lowers energy and operating costs. The infrastructure is stable, fast, secure, ready to scale and aligned with industry best practices.
Planning to build or expand your hosting infrastructure? Let’s talk about your project.
It is an environment in which every critical component — network, servers, storage — has a standby. If a switch, server or storage controller fails, services keep running because another component takes over.
Proxmox VE is a mature open-source platform with built-in HA clustering, ZFS support and backups through Proxmox Backup Server. It has no per-core licensing, which significantly reduces costs at thousands of cores.
ZFS protects data integrity with checksums and offers snapshots, replication, compression and deduplication. On NVMe drives it gives very fast access to customer data.
Yes. The three-layer architecture lets you independently add servers to the Proxmox cluster, more ZFS arrays and ports on the 100 GbE network.
Yes. In this project we moved customers’ hosting, VPS and dedicated environments onto the new infrastructure, and after go-live took over its administration and maintenance.
Yes. We delivered this project for a hosting provider in the Czech Republic. We design, deploy and operate infrastructure for clients across the European Union.
Other projects
Private cloud for a process platform, digital patient onboarding, SLA-driven case handling and ReconMore vulnerability monitoring for a medical clinic network.
Operational audit, a low-code process platform in a private cloud and ReconMore vulnerability monitoring for a high-volume telemedicine organization.
Purchase approval workflows, complaint handling with SLAs, electronic CAPA/HACCP records and a single unified environment maintained by Remote Admin.