Re: Tunnel MTU

[Mohacsi Janos <mohacsi@niif.hu>]

On Tue, 28 Jul 2009, Iljitsch van Beijnum wrote:

> On 28 jul 2009, at 19:44, Mark Townsley wrote:
>
> > I'd like to clarify of course that some of the MTU issues we discussed were 
> > not specific to tunneling, but to mismatched MTUs on a home LAN vs. WAN 
> > interface in general.
>
> In IPv6 you can easily broadcast your WAN MTU in RAs on the LAN. So if your 
> WAN has 1337 you simply have an MTU option with "1337" in RAs that are sent 
> out on the LAN side.
>
> > The extra 20 bytes of a 6rd or 6to4
>
> With 6to4 most implementations simply use 1280.
>
> > encapsulation isn't significant when trying to solve support of 9K jumbo 
> > frames and standard 1500 byte ethernet MTUs in the same network.
>
> I have an expired draft that solves exactly that issue:
>
> http://tools.ietf.org/html/draft-van-beijnum-multi-mtu-02
>
> I intend to bring it back to life in a much improved incarnation at some 
> point, didn't get that done before this meeting.

I somehow lost the in the discusion:

According the path-MTU discovery process should work this way: the CPE 
(and subsequent routers) should notifiy the sender (with ICMPv6 packet too 
big message) to reduce MTU if the sender wants to send something over the 
WAN links (including tunnels).

So actual path-MTU may differ by destinations.

Actually the problem can be find out the sensible MTU for tunnels? Are we 
discussing the this value?

Or we are discussing something better than PMTUD?

Best Regards,
		Janos