Re: Dialup over Ethernet

On 14/09/17 09:33, Peter Lambrechtsen wrote:
> For me the restore time in the event of loss of service to customers is 
> the main reason why PPPoE is superior to DHCP/IPoE.

Thanks for all the details.

I do remain puzzled by the use of a stateful protocol (PPPoE) with a 
timeout (LCP keepalive) leading to better restore time than a mostly 
stateless approach (eg, VRRP for the gateway address, let CPEs keep 
using their DHCP lease IP, separate DHCP servers with state, minimal 
state in customer-facing gateway other than VRRP-master-of-the-second -- 
which is a common approach in, eg, datacentre networking).

> Voice ATA services on the same handover can take up to 30 mins to
> restore even though the DHCP lease is set down to 5 mins.

Presumably this is until they give up on trying to renew their old lease 
and get around to asking for a different one.  But it's not clear why 
they need a new IP anyway, at least over a short break like a head end 
device reboot.  If the IP/subnet/gateway stays the same, IIRC most DHCP 
clients will successfully re-request it from another DHCP server on the 
same subnet if their original DHCP server goes away.

> Another advantage is to find out the customers current sync rate on 
> copper as the vast majority of Chorus EUBA/WVS DLSAMs send you the 
> connect rate information in the PADI

That's definitely an argument for PPPoE on xDSL for legacy reasons; but 
UFB isn't over copper...

> Even in the UFB world I would always expect some sort of router / 
> Customer RGW between the ONT and the customers devices. I wouldn’t 
> expect a customer to plug in a switch into an ONT then have multiple 
> devices sending DHCP requests to the BNG

Nor would I, FWIW,  I'd expect there to be some sort of router/CPE/RGW 
device at the edge.  The usual way to handle that in a 
just-switched-Ethernet world is to put a MAC limit on the port; it seems 
like an obvious thing that an ONT could enforce (eg, 1 MAC, dynamically 
learnt, forgotten as soon as ethernet carrier on link to CPE goes away 
so end user can change CPE at will).

> Since unconvinced to see the compelling reason case why IPoE/DHCP is any 
> better than a PPPoE link, as I can only see downsides to IPoE.

Definitely seems to be a religious war :-)

Thank you all, I'm much better informed as to why we ISPs are still 
doing PPPoE in a UFB/metro-ethernet environment.  It still doesn't seem 
like a good tradeoff to me, but at least y'all are getting more than 
username/password authentication out of it!

Ewen

_______________________________________________
NZNOG mailing list
[email protected]
https://list.waikato.ac.nz/mailman/listinfo/nznog