I've just started with broadband and I found that with pppoe, pppd takes about five minutes to make the connection, during which there are numerous timeouts waiting for PADO packets. Once the connection's up though, it works just fine.

I asked my dsl provider (enta.net) if they had any advice about this and was told they don't use pppoe, they use pppoa. So I reset my adsl modem/router to pppoa, dug out a couple of pppoa scripts from /usr/share/doc/ppp/examples, edited in my username and tried it that way. No joy: the ppp daemon died almost as soon as it was born. Well, I prefer a connection that takes five minutes to one that doesn't come up at all, so I went back to pppoe and I'm still using it.

What I'd like to know is how pppoe can work at all in this kind of situation. If the DSLAM at the exchange only talks pppoa, how does it manage to talk pppoe to my router? Because a PADO packet does eventually arrive with a hardware address that must belong to the DSLAM (because it certainly doesn't belong to any of the interfaces on the router). Immediately afterwards I receive IP and DNS addresses as well and everything is tickety-boo.

The enginner at enta.net support seemed to think that the whole PADO thing is pppoe-specific but I think he must be wrong about that. I'd welcome any clarification or a useful link.