[c-nsp] MTU issues with tunnel switching

[Clayton Zekelman]

J.P. is referring to the model where Bell Canada uses a router (Juniper ERX-1400) to act as a PPPoE LAC, forwarding sessions that come in as PPPoE to wholesale customers over L2TP tunnels.

The additional per packet overhead is introduced in the L2TP tunnel - 40 bytes per packet.  The fact that the session is entering the Bell router as a PPPoE session has nothing to do with the overhead on the L2TP tunnel.

The real problem has nothing to do with PPPoE, but everything to do with doing L2TP over FastEthernet - where you have to fit the L2TP packet along with 40 bytes of overhead on a physical medium (FastE) that only allows packets with a MTU of 1500 bytes.



----- Original Message ---------------

Subject: Re: [c-nsp] MTU issues with tunnel switching
   From: Dennis Peng <dpeng at cisco.com>
   Date: Mon, 28 Mar 2005 14:43:36 -0800
     To: "J.P. Racine" <racinejp at vianet.ca>
     Cc: cisco-nsp at puck.nether.net

>J.P. Racine [racinejp at vianet.ca] wrote:
>> Could someone confirm that if I apply tcp adjust-mss on a virtual template
>> for a VPDN group that it won't adjust the size if the packets original MSS
>> packet header value if is already smaller than the set value in the
>> configured template?  
>
>Yes, definitely.
> 
>> Does running PPPeoE over L2TP create more overhead than other types of PPPoE
>> implementations?
>
>I'm not sure what you mean by this question, could you clarify?
>
>Dennis
> 
>> Regards,
>> 
>> J.P. Racine
>> racinejp at vianet.ca
>> 
>> 
>> 
>> _______________________________________________
>> cisco-nsp mailing list  cisco-nsp at puck.nether.net
>> https://puck.nether.net/mailman/listinfo/cisco-nsp
>> archive at http://puck.nether.net/pipermail/cisco-nsp/
>_______________________________________________
>cisco-nsp mailing list  cisco-nsp at puck.nether.net
>https://puck.nether.net/mailman/listinfo/cisco-nsp
>archive at http://puck.nether.net/pipermail/cisco-nsp/