Interface naming across vendors

Interface naming is trivial until you work across vendors at speed, at which point it becomes a genuine source of error — typing a correct command against the wrong interface.

The conventions

Cisco names by speed and position: GigabitEthernet0/1, TenGigabitEthernet1/0/3. The abbreviation is universal in practice: Gi0/1, Te1/0/3. Numbering typically starts at 1 within a module.

Juniper encodes location: ge-0/0/1 is a gigabit interface in slot 0, PIC 0, port 1. Then .0 for the logical unit, which is mandatory for configuration. Numbering starts at 0.

MikroTik names by function and order: ether1, sfp-sfpplus1, wlan1. Interfaces can be renamed freely, which is either helpful or confusing depending on whether anyone documented it.

Linux uses predictable names derived from bus position: enp3s0, ens160. Ugly, but stable across reboots and hardware changes, which the old eth0 naming was not.

The off-by-one

Juniper and Linux start at 0; Cisco switches generally start at 1, but Cisco routers start at 0 (Gi0/0/0). The first physical port on the front panel might be ge-0/0/0 or Gi0/1. Moving between platforms in one session, this catches people, and the command succeeds against the wrong interface rather than failing.

Logical units

Juniper’s unit 0 has no direct Cisco equivalent for a simple interface, and omitting it is a common early mistake. It exists so one physical interface can carry many logical ones, which is what makes VLAN subinterfaces uniform rather than a special case.

The practical defence

Describe every interface, on every platform, with what it connects to. Then you verify against the description rather than the number:

show interface description
show interfaces descriptions        # Junos
/interface print detail             # RouterOS comments

Before any disruptive command, confirm the description matches what you intend to affect. It takes seconds and it is the difference between shutting the right port and shutting a production link.