Re: [L2tpext] Offset size

W. Mark Townsley wrote:

> If you have an odd-sized layer two frame being carried between your tunnel
> header and the framed network packets (e.g. an IP packet being framed by a PPP
> packet with PFC negotiated, resulting in a 3 byte PPP frame) then it could help
> the performance of some implementations to insert padding bytes to ensure
> alignment of the framed network packet (e.g. the IP packet). 

OK in that case the offset can not exceed 3 bytes to align the end of 
the header on the end of a 32-bits word?
So why the RFC propose a 2 bytes long field.

> At one time, it's affect on at least some implementations was measurable,
> particularly on an LNS which routed the tunneled, framed, IP packet locally
> rather than sending it out another physical connection. The implications today
> may not be as marked, and is in fact becoming more and more contraversial as to
> whether it is even necessary, but still could theoretically be an issue. It is
> always, of course, optional.

In L2TP data packet, it seems that optional fields are not used. Is that 
the current implementation?

> 
> 
> - Mark
> 
> Cédric Fau wrote:
> 
>> Hello,
>> 
>> I have some issues with the offset size field. I don't understand its aim.
>> If it is still present in the L2TP header in the
>> draft-ietf-l2tpext-l2tp-ppp-01.txt, that's
>> because there is a use of this field. But what kind of use ?
>> 
>> Cédric Fau
>> mailto: cedric.fau@etu.enseeiht.fr
>> 
>> Student in Telecommunication and Network Engineering
>> http://www.enseeiht.fr/
>> 
>> _______________________________________________
>> L2tpext mailing list
>> L2tpext@ietf.org
>> https://www1.ietf.org/mailman/listinfo/l2tpext
> 

_______________________________________________
L2tpext mailing list
L2tpext@ietf.org
https://www1.ietf.org/mailman/listinfo/l2tpext