Originally Posted by ToddAndMargo
Thank you! Now I understand. New naming convention that is suppose to make it easier to figure out what Ethernet port is what. Not sure how it is going to help, but what matters is that now I understand what is going on.
For a typical desktop user where there is only one Ethernet port, it shouldn't make any difference. It's just a name.
Where it helps is when you've got a system with lots of ports on cards (like a big server.)
If a card fails and you swap it out for a new one, the port name shouldn't change. With the previous scheme, there was no guarantee. When a card gets swapped, the MAC address changes (so the HWADDR in the config file is useless) and it may use a different device driver (so the load-order may change.)
If you swap only one card, software could probably infer what to do by seeing that all but one config file has a HWADDR that matches hardware. But what if you're swapping multiple cards or if you're swapping a multi-port card? With the old scheme, there's no good way to avoid making the operator go through a time-consuming process of validating (and maybe re-assigning) device names after the upgrade.