The name of a device class acts as an identifier for this class. It simply is a string and can be chosen mostly freely. The name can not contain the character /, as it is internally used to separate inherited device classes. It also needs to be unique among all devices classes which inherit from the same parent device classes.

When creating a top level device class, which is a device class that directly inherits from generic, it is best practise to name the device class after the operating system or network device vendor it represents. When creating a sub device class which specifies an already existing class for some devices with different behavior, the name of that device class should show which devices of its parent device class it is responsible for.

The name of this sub device class does not need to include any information about its parent class, as the name of a classes parent is automatically added to it internally (except for generic). For example, ceraos has a sub device class ip-10. Internally this device class is called ceraos/ip-10.