What is colocation?

Last updated August 27, 2026 · Reviewed by James San Filippo

Colocation is renting space, power, cooling and connectivity in a data center for servers you own and manage yourself. The provider supplies the building and everything that keeps it running; you keep the hardware, the operating systems and the data. Colo Solutions has provided colocation in downtown Orlando since 1997, in cabinets from one-third rack (12RU) to full rack (42RU) and private cages.

The short version

Colocation — often shortened to “colo” — means putting your equipment in someone else’s building.

You buy and own the servers. The provider supplies the rack space, conditioned and backed-up power, cooling, physical security and network access. You keep administrative control of everything running on the hardware; the provider never logs into it.

That boundary is the defining feature, and it is what separates colocation from both cloud and managed IT.

How it works in practice

You rent by the cabinet, or a portion of one. A one-third rack gives you 12 rack units; a full rack gives 42. Larger deployments take a private cage.

Your equipment is delivered and racked, connected to power — usually two feeds, from separate paths — and to whichever network provider you choose. From then on you administer it remotely, and visit when hardware needs hands.

The provider is responsible for keeping the environment within specification: power available, temperature and humidity controlled, and the building secure.

What a colocation data center provides

• Redundant power. Utility feeds, UPS to carry the gap, and generators for extended outages. Colo Solutions runs two independent utility feeds, one on the Priority 1 hospital grid, with three diesel generators in an N+1 farm.
• Purpose-built cooling, with humidity control, monitored continuously.
• Physical security. Dual-factor access, a mantrap, uniquely keyed cabinets and recorded video.
• Connectivity. In a carrier-neutral facility, a choice of carriers reachable by cross connect rather than a single provider you are stuck with.
• Fire detection and suppression, plus water and environmental monitoring.

Understanding data center tiers — and a common error

Facilities are often described using the Uptime Institute’s four tiers. These are routinely misquoted as describing how long a site survives a power outage. They do not. The tiers describe redundancy and maintainability:

• Tier I — Basic capacity. Single path for power and cooling, no redundant components. Maintenance means shutting down.
• Tier II — Redundant capacity components. Redundant components, still a single distribution path.
• Tier III — Concurrently maintainable. Multiple distribution paths, one active. Any component can be taken out of service for maintenance without shutting down IT equipment.
• Tier IV — Fault tolerant. Multiple independent, physically isolated systems. Any single failure occurs without impacting IT equipment.

Two things follow. A tier is a certification awarded to a specific facility, so “Tier 3 design” and “Tier 3 certified” are different claims — ask which is meant. And a higher tier is not automatically better for you: it costs more, and what most tenants need is redundancy across the components that fail in practice, which is why many facilities, Colo Solutions included, are built and described as N+1.

Colocation versus the alternatives

Versus a server room in your office: you get infrastructure you would not build for one company’s equipment, and carrier choice you cannot get on a single commercial street.

Versus public cloud: you keep the hardware and the control, and the cost is predictable rather than usage-metered. You also keep the responsibility — nobody else is patching those machines.

Versus managed hosting: in colocation the equipment is yours to administer. If you want somebody else accountable for the operating system, that is a different service.