Skip to content

Equinix Metal Servers

Deploying servers at Equinix Metalâ„¢ is very similar to working with compute instances at AWS, GCP, Digital Ocean and others. However, since Equinix Metal is 100% focused on physical bare metal servers, there are some differences. This is especially true around workload portability, networking setup, and optimization.

A catalog of Equinix Metal servers is available on the Servers product page of the Equinix Metal website.

Server Naming Convention

All instances have a name that follows a formula to help you understand its purpose and broad specifications: class + generation + size + architecture (optional). By piecing these features together, servers get names such as c3.medium.x86, indicating a medium compute class server, in its 2nd generation, with an x86 processor.


Classes are groupings derived from common use cases. If you're an AWS user, this will seem pretty familiar. Equinix Metal organizes servers into the following classes:

  • Compute (c) - Compute focused, with a modest RAM footprint.
  • Memory (m) - Memory heavy, with a generous RAM to core ratio.
  • Network (n) - Focused on network use cases, such as ingress / load balancing.
  • Storage (s) - Scale out boxes for storage scenarios.


Generations are updated when Equinix Metal significantly upgrades or changes server components. Usually a new generation is released around a major processor refresh, but often it also relates to the underlying hardware configuration or chassis. A 2nd generation system is not necessarily outdated, it is just the second iteration of a particular configuration.


The server size is a label for the total amount of resources across CPU, memory, and disk. Generally, the larger the server size the more resources it has and the more expensive it is. Equinix Metal server sizes currently include small, medium, large, and xlarge.


The architecture is the server's processor. x86 indicates the server has a x86 processor and arm is used for our ARM processor servers.

Servers in the API

Information about Equinix Metal servers in the API can be found in the plans and devices endpoints.


The plans API endpoints generally return information about what types of servers your organization, project, and user account can provision.

Sending a GET request to the /plans endpoint will return a list of servers that your user account is able to provision with lots of information about each, including name, a description, and which data centers the servers are available in.

curl -X GET -H 'X-Auth-Token: <API_TOKEN>'

Sending a GET request to the /projects/{id}/plans endpoint, where the id is the Project ID, will return a list of servers that are available to that project.

curl -X GET -H 'X-Auth-Token: <API_TOKEN>'{id}/plans

Ans sending a GET request to the /organizations/{id}/plans endpoint, where the id is the Organization ID, will return a list servers that available to your organization.

curl -X GET -H 'X-Auth-Token: <API_TOKEN>'{id}/plans


The devices API endpoints are the endpoints that provision and manage servers.

For example, to retrieve information about all the servers that you have provisioned in a project, send a GET request to the /projects/{id}/devices endpoint.

curl -X GET -H 'X-Auth-Token: <API_TOKEN>'{id}/devices

You can provision servers with a POST to the /projects/{id}/devices endpoint. And a more in-depth explanation of on-demand server provisioning is on the On-Demand Servers page.

There are many more features that the Equinix Metal API offers for server management and configuration, please see the API reference.