Give demanding work the whole machine.

When a dedicated server makes sense

A dedicated machine removes other customer virtual machines from your host. It can provide more consistent resources and lets you choose a hypervisor, storage layout and host-level tuning. You still share upstream connectivity and the facilities that power and cool the equipment.

Build the right storage layout

Two drives are not automatically a backup. Mirroring can improve continuity when one drive fails; striping can favour capacity or throughput while increasing exposure to a drive failure. Keep independent backups and test a restore before production data depends on the machine.

• Use encryption where the workload requires it
• Watch SMART and filesystem health where available
• Document partitioning and recovery steps

Remote control needs strong boundaries

Out-of-band management is powerful because it can operate below the installed operating system. Restrict management access, rotate credentials and use it only from trusted networks. Availability and access method must be confirmed for the selected configuration.

A machine matched to the workload

Compare the range from entry infrastructure through high-density compute. Confirm the hardware, stock, installation time, replacement procedure and network terms for your selected order.

Engineering options beyond the five standard machines

GPU nodes, additional NVMe storage and specific network interface cards require a custom-configuration request. Their price, stock, compatibility, power and lead time require an individual verified quote. Confirm stock and the full quote before planning a workload around custom hardware.

Monthly or annual hardware billing

All five dedicated tiers offer monthly billing or an upfront payment for one 12-month period with a 30% discount compared with 12 separate monthly payments. The annual total is 12 times the listed monthly price multiplied by 0.70, rounded once to cents, with no setup fee. Check renewal, cancellation and refund terms with your order. Review remote IPMI/KVM access and operating-system reinstall options with your order. The catalogue targets roughly five minutes for a reinstall and thirty minutes for initial deployment; confirm the delivery terms for the selected machine.

What happens when a part fails?

The hardware-support agreement should explain component replacement or storage migration after a confirmed fault. Check the restoration window, data-risk limits, spare availability and customer responsibilities. Confirm component availability and the applicable replacement commitment for your selected hardware.

What belongs to the dedicated service

The dedicated range reserves a complete physical machine for one customer. The specified processor, ECC memory and NVMe drives belong to that configuration; there is no customer VM hypervisor inserted by the catalogue. Network links, transit and facility systems remain shared services. An included management console does not turn the machine into a managed application service.

• IPMI or KVM access for console and recovery tasks
• Operating-system reinstalls without a separate reinstall charge
• IPv4 and IPv6 connectivity, edge mitigation and a technical support path
• No setup fee in the published catalogue

Reinstall deliberately, then verify access

A reinstall replaces the installed system and can erase application data. First export anything that must survive, confirm the chosen disks and image, and keep the management console available. The intended reinstall target is around five minutes, separate from the roughly thirty-minute initial dedicated provisioning target. IPMI/KVM access is intended to be available immediately when the machine is activated, rather than after a separate support request. Actual completion must be reported by the provisioning service, not inferred from a countdown.

• Save an independent backup before confirming
• Check the new SSH host key through a trusted channel
• Apply updates, recreate the firewall and test a permitted administrator login

A hardware SLA needs a hardware recovery process

A hardware fault and an application outage need different responses. A useful hardware agreement identifies the component checks, replacement or migration procedure, customer notifications and the point at which recovery is considered complete. The hardware terms for the selected order govern any replacement window, recovery procedure and service-credit remedy. Check that scope before relying on a particular repair deadline.

Custom GPU, storage and network hardware

The five standard configurations are a starting catalogue. A different GPU, additional NVMe devices or a particular NIC requires a separate engineering specification and quote. State the application, accelerator memory, disk count and capacity, endurance needs, interface speed and expected sustained load. Compatibility, power, cooling, delivery and support scope must be confirmed before a custom order is priced.

• A GPU request does not mean a GPU is included in a standard tier
• Additional IP quantities, RAID layout and extra hardware fees require a written quote
• Use the contact information page to find the verified enquiry channel when the service opens