Re: [v4v6interim] Fragmentation Options in NAT6 presentation

On Monday 06 October 2008 19:34:51 ext Wes Beebee (wbeebee), you wrote:
> I do agree that RFC 4821 is the right solution, and should be used in
> recommendations in IETF documents.
>
> However, how many stacks actually implement RFC 4821?

There are really two separate categories:

For TCP (and SCTP), this needs to be supported by the IP stack. The widespread 
use of TCP MSS clamping makes it obvious that many if not most IP stacks do 
not support this as yet.

For UDP, this needs to be implemented directly into the application-layer, 
with some additional support from the IP stack (to allow sending too large 
packets with DF bit). As yet, Linux is the only IP stack that explicitly 
supports this via some "proprietary" socket API options. As for applications, 
it is difficult to gather meaningful data. But considering the lack of 
support from lower layers, it is likely that this is mostly not implemented. 
I empirically believe that most applications simply have a "safe" hard-coded 
MTU for UDP, e.g. 1400 bytes or whatever.

-- 
Rémi Denis-Courmont
Maemo Software, Nokia Devices R&D
_______________________________________________
v4v6interim mailing list
v4v6interim@ietf.org
https://www.ietf.org/mailman/listinfo/v4v6interim