A Linux dedicated server is bare-metal hardware (single-tenant physical machine) rented from a hosting provider. The team gets the entire CPU, RAM, disk, and network - no hypervisor, no noisy neighbors, no shared kernel. The team that runs latency-sensitive workloads, single-tenant databases, or large memory-footprint apps picks dedicated. The team that runs short-lived cloud VMs or stateless APIs picks VPS or cloud.
Table of contents
- What ‘dedicated’ actually means
- When dedicated is the right pick
- When to pick something else
- Dedicated hosting providers
- Operating a dedicated server
- FAQ
What ‘dedicated’ actually means
A dedicated server is a physical machine that the provider rents to a single customer. The hardware - CPU, RAM, disk, NIC - belongs to the customer’s workload alone. There is no hypervisor layer; the customer runs their OS directly on the metal.
Compare:
- Dedicated: Physical hardware, single tenant, full hardware access. No noisy neighbors.
- VPS / Cloud VM: Virtual machine on shared hardware, hypervisor-isolated. Slight overhead and noisy neighbor risk.
- Managed Kubernetes: Cluster of VMs (cloud) or bare metal (on-prem), with orchestration.
- Colocation: Customer owns the hardware, hosts it in a provider’s data center.
The team that picks the right tier matches the workload to the isolation level. The team that picks dedicated for everything overpays for noise isolation they don’t need.
When dedicated is the right pick
Dedicated makes sense when:
- The workload is single-threaded and CPU-bound (noisy neighbors on shared hardware cause variance).
- The workload needs lots of RAM (>256GB) - cloud VMs get expensive at this scale.
- The workload needs bare-metal features: SR-IOV, DPDK, custom kernel modules, large pages, NUMA pinning.
- Compliance requires single-tenant hardware (some financial and healthcare workloads).
- Cost: dedicated hardware is cheaper per-unit than equivalent cloud at scale.
The team that has these requirements has the right case for dedicated. The team that runs a small web app does not.
When to pick something else
VPS / cloud VM makes sense when:
- The workload is elastic (needs to scale up and down).
- The team does not want to manage hardware.
- The team’s geographic coverage needs are global (cloud has more regions than dedicated providers).
Managed Kubernetes makes sense when:
- The team runs many services and needs orchestration.
- The team wants to focus on application code, not infrastructure.
- The team has variable load and needs horizontal scaling.
The team that picks the right tool matches the operational model to the team’s skills.
Dedicated hosting providers
Common dedicated providers:
- Hetzner (Germany, Finland, US) - good price/performance, popular in EU.
- OVH (France, Canada, US) - global coverage, bare-metal and cloud.
- Equinix / Digital Realty - colocation and managed bare metal.
- SingleHop, LiquidWeb, HostGator - US-centric, less common in 2026.
- IBM, Dell, Supermicro direct - for colocation / on-prem.
The team that picks a provider looks at: hardware specs, network quality (bandwidth and peering), datacenter location, support SLA, and price.
Operating a dedicated server
The team that runs dedicated owns:
- OS installation and patching.
- Security hardening (firewall, fail2ban, SSH config).
- Backups (provider’s snapshot, offsite, or both).
- Monitoring (CPU, RAM, disk, network).
- Hardware failure replacement (provider does the swap, but you coordinate).
The team that uses a managed dedicated provider offloads some of these. The team that self-manages has full control but more work.
A typical setup: Hetzner or OVH dedicated box, Ubuntu LTS, Ansible for config management, Prometheus + node_exporter for monitoring, restic to Backblaze B2 for backups.
FAQ
How much does a dedicated server cost?
Mid-range (AMD EPYC or Xeon, 64GB RAM, 2x1TB NVMe, 1Gbps unmetered): $80-150/month at Hetzner, $150-300 at OVH, more at US providers. The team that prices dedicated against equivalent cloud VMs at >50% utilization usually finds dedicated cheaper.
Can I run Docker on a dedicated server?
Yes. Docker runs on bare metal. Many dedicated setups run a small Kubernetes cluster (k3s, microk8s) on dedicated hardware for the cloud-like orchestration without the cloud cost.
Is dedicated more secure than cloud?
Single-tenant hardware reduces some attack surface (no hypervisor escapes from neighbors). But cloud providers invest heavily in physical security, network security, and patching. The team that picks dedicated for security alone usually finds that operational discipline matters more than the isolation level.
What about serverless and edge computing?
Serverless (Lambda, Cloudflare Workers) and edge platforms (Fly.io, Render) cover some workloads that previously needed a dedicated server. The team that picks serverless for short-lived, low-latency workloads, dedicated for stateful single-tenant apps, and cloud VMs for elastic workloads has the right mental model.
Can I migrate from dedicated to cloud or vice versa?
Yes. Standard Linux tooling (rsync, tar, dd) works for migration. The main challenges are network (the team that migrates large databases needs to plan bandwidth) and downtime (most migrations need a maintenance window).
Is a dedicated server the same as bare metal?
In hosting, yes - both mean single-tenant physical hardware. The cloud provider ‘bare metal’ offerings (AWS i3.metal, GCP bare-metal) are dedicated servers with cloud-style APIs and billing.
If you are sizing the infrastructure for the kind of project this post covers, the RunxBuild hosting calculator is the right place to model the line items. The compute, the memory, the storage, the bandwidth, the database - each one is a separate number, and the team’s mental model for the platform is the sum of those numbers. The RunxBuild dashboard is where the team sees the actual usage in one place.
Useful related references: