Cagelab
Guide

What is Colocation?

Colocation means placing your own servers and infrastructure in a professionally managed data centre, and paying the facility operator for power, space, cooling and connectivity.

The basic model

In a colocation arrangement, you own the servers, switches, storage arrays and other hardware. You deliver them to a facility operated by a data centre company, which provides the physical space (measured in rack units, half racks or full racks), electrical power (measured in kilowatts), cooling, physical security, fire suppression, and one or more internet connectivity options.

The division of responsibility is usually clear: the facility is responsible for everything up to the front of your rack. You are responsible for everything inside it. This is sometimes described as the shared responsibility model in the colocation context, though the term is more commonly associated with cloud computing.

This model gives you the benefits of enterprise-grade infrastructure (redundant power, precision cooling, physical security, diverse connectivity) without having to build and maintain it yourself. The capital cost of constructing and fitting out a data centre facility is prohibitive for most organisations; colocation allows you to use that infrastructure for a monthly fee.

What you pay for

Colocation pricing typically has several components. The primary cost is the power allocation: the contracted amount of electrical power reserved for your hardware, usually priced per kilowatt per month. Space is often included within the power pricing for standard density deployments, though high-density or custom configurations may price space separately.

In addition to power and space, you will typically pay for connectivity: a physical connection from your rack to the facility's network infrastructure and, from there, to the internet or to specific carriers and cloud providers. Connectivity pricing varies significantly by bandwidth, redundancy and carrier choice.

Additional services such as remote hands (on-site technician support), cross-connects (dedicated physical connections to other tenants or providers in the facility), and managed services are priced separately and vary by facility and provider.

When colocation makes sense

Colocation is generally the right choice when your hardware base is stable and predictable, your workloads have consistent performance requirements, and you have the IT capability to manage your own hardware. It typically becomes financially competitive with public cloud at sustained workloads above 5 to 10 kW, where the three-year total cost of cloud compute often exceeds the equivalent colocation cost by a meaningful margin.

Use the Colocation vs Cloud Calculator to model the three-year comparison for your specific workload, or read the colocation vs cloud guide for a more detailed analysis.

Key terms explained

Rack unit (U): The standard unit for measuring vertical space in a server rack. A standard rack is 42U. A 1U server occupies one rack unit of height.

kW (kilowatt): The unit used to measure power allocation and power density. Most standard colocation contracts are priced per kW of contracted power. A typical server rack draws 5 to 15 kW; GPU racks can draw 30 to 80 kW or more.

PUE (Power Usage Effectiveness): A measure of data centre energy efficiency. A PUE of 1.0 would mean all power consumed goes to IT equipment. A typical modern facility runs at 1.2 to 1.5. Lower is better.

Redundancy: The duplication of critical systems (power, cooling, connectivity) to ensure availability if one component fails. N+1 means one additional component above the minimum required. 2N means fully duplicated across two independent paths.

Frequently asked questions

What is the difference between colocation and a managed hosting service?

In colocation, you own the servers and hardware. You place them in a facility owned and operated by a data centre provider, who supplies the power, cooling, physical security and connectivity. In a managed hosting service, the provider owns the hardware as well and you rent compute capacity. Colocation gives you more control and typically lower long-term costs for stable workloads, but requires you to manage the hardware.

How is colocation priced?

Colocation is typically priced per rack, half rack or quarter rack per month, or on a per-kilowatt basis for high-density deployments. The monthly charge includes a power allocation (usually measured in kW or kVA), basic connectivity, and physical security. Additional services such as remote hands, extra connectivity, cross-connects and managed services are priced separately.

What is a cross-connect in colocation?

A cross-connect is a dedicated physical cable connecting your cage or rack directly to another tenant or carrier within the same facility, without the traffic crossing the public internet. Cross-connects are used to connect to cloud providers, carriers, or internet exchanges that have a point of presence in the facility.

What does remote hands mean?

Remote hands is a service provided by the colocation facility where their on-site technicians perform physical tasks in your cage or rack on your behalf. This includes rebooting servers, swapping cables, inserting media, and similar tasks. It is billed at an hourly rate and is useful for organisations that do not have IT staff physically near the facility.

What is a Tier 3 data centre?

The Uptime Institute Tier classification system rates data centre infrastructure resilience from Tier 1 (basic) to Tier 4 (fault tolerant). Tier 3 is the most common enterprise standard, requiring N+1 redundancy on all critical systems (power, cooling, connectivity) and 99.982% availability. Tier 3 facilities can be maintained without taking the facility offline.

Evaluating colocation for your organisation?

Use the free Colocation vs Cloud Calculator to model your three-year costs, or get in touch with Cagelab for help identifying the right UK providers for your workload.