The price of server colocation is based on 3 main factors:
-
The space used in the rack, expressed in rack units (U)
-
The electrical consumption of the hosted infrastructure, measured in Amperes
-
The bandwidth used (and its type), as well as the number of public IPv4 addresses required
-
The balance between the different resources requested
1) Rack Space Used
Your server must be 19" rack-mountable and come with its own mounting rails.
The rental fee is defined by the height of your server, which determines how many rack units it occupies.
A rack unit (symbol: U) is the standardized height of electronic equipment designed to be installed in a rack.
One rack unit measures 1.75 inches (44.45 mm) in height.
Depending on the datacenter selected, the price per rack unit may vary.
2) Power Consumption of Your Machines
One of the key factors is the actual power consumption of your server.
To know this precisely, you should plug your server into a power outlet that displays real-time consumption.
Do not rely on the power rating written on the power supplies — that number represents the maximum output the PSU can deliver, but actual server consumption is always lower.
Be cautious with servers that have redundant power supplies. On some systems, simply connecting a second PSU can increase power draw by 70 Watts (0.3 Amps).
3) Bandwidth Usage
We can provide several types of bandwidth/connectivity options:
- Volume-based bandwidth on a 1 Gbps port
In this case, the total data volume transferred through the network port is measured, and you are billed based on the agreed price per GB of data.
This is generally the most expensive method per GB, but it allows you to have 1 Gbps connectivity for very small data volumes.
- Flat-fee bandwidth on 100 Mbps, 1 Gbps, or 10 Gbps ports
In this case, your server is connected to an interface with a fixed speed limit, and you pay a flat monthly fee.
This is the most cost-effective option because we can plan ahead for the client’s usage and do not need to monitor data volume.
It is ideal if your traffic usage is known in advance. We can offer flat-fee bandwidth in 100 Mbps increments.
- 95th percentile billing
In this model, we provide a port with a specified capacity and charge based on your 95th percentile bandwidth usage.
This allows you to handle large traffic spikes without being penalized. For example, if your average usage is 90 Mbps with occasional peaks up to 500 Mbps, you won't be billed for those short spikes.
IPv4 Addresses
The number of public IPv4 addresses requested also affects pricing.
IPv4 addresses are becoming increasingly expensive, so a monthly rental fee is charged for each one.
However, each client is entitled to a /48 IPv6 range upon request (with justification).
4) Balance Between Requested Resources
We always provide a custom quote based on the balance of resources requested by the client.
For example, a 1U server drawing 10 Amps and using only 1 Mbps of bandwidth has a greater impact on our infrastructure than a 1U server using 0.5 Amps and 1000 Mbps of flat-fee bandwidth.
Colocation services are often significantly more expensive than renting a dedicated server, due to several factors:
-
Dedicated servers are standardized, allowing optimal use of datacenter rack space
-
Bandwidth for dedicated servers is shared among clients, and most customers use less than 10% of their allocated bandwidth. This oversubscription reduces costs.
In contrast, colocation clients are usually aware of their actual usage, and the allocated bandwidth is genuinely consumed.
