option 212 for 6RD

[Jeroen Massar]

On 2013-01-15 06:50, Mikael Abrahamsson wrote:
> On Mon, 14 Jan 2013, Jeroen Massar wrote:
> 
>> As the tunnel should indeed be 1480 or lower, but the LAN can just
>> stay at 1500 and changing that would cause all kinds of other odd
>> issues I would think for hosts that don't take the info from the RA...
>> and the CPE should really be doing PTBs thus lowering the MTU is not
>> needed as it will send a PTB because of the tunnel interface having
>> the lower MTU (eg 1480).
> 
> This causes almost every destination to have to do PMTUD (6RD gateway
> sending PTB).

But the first hop is local likely at 100mbit or GigE thus that response
comes back very quickly from that one. Also note that the destination
caches will retain this information for quite a bit.

> I see no reason to keep LAN at 1500 if the tunnel is 1480.

What about traffic between hosts on the same LAN? (see below too)

> "all" traffic
> will need to do PTB causing unneccessary delay for initial connections
> to most if not all destinations (web server sending 1500 packet, getting
> PTB from 6RD gateway, then sending 1480 packet).

Except for TCP's MSS there is no way of universally indicating a minimal
MTU to a remote server, thus indeed if there are one (or more) changes
in MTU on the path you will be resending those packets and getting PTBs
for them the first time you talk to that server.

There is no way around that, though in the sole case of TCP indeed
setting the MSS works. Well there is a 'work around' and that is the one
you are proposing: make everything 1280 as then you are sure that it is
all okay..... but that is flawed.

Note also that setting your local network to a wrong MTU does not
resolve this. Typically packets from the server will be large while the
client only sends small one, and you are only controlling the sending of
packets with your local MTU, not what the server is sending, which will
in most cases happily be at 1500 and maybe even jumbo.

> Personally I have run my home with an IPv6 MTU of 1400 for several
> years.

IMHO it is very wrong do so and you are only hurting your own network.

Yes, it works, but why do that if there is a known working solution:
working PMTU.

> We had this discussion in IETF v6ops as well, but then it was
> mobile.

(I must have missed it....)

> The suggestion there was to advertise the same MTU on the LAN as
> was seen in RAs on the device<->mobile network GTP tunnel. This makes a
> lot of sense to me and I strongly supported it.

I wonder who made that suggestion and with what kind of reasoning as it
does not help the typical client(x)<->tunnel(<1500))<->server(1500) case
at all as the server will try to send an initial packet at 1500 anyway
when it is sending large content back to the client.

> What's your reason to so much want 1500 on the LAN that you'd rather
> take that than the delay for most connection I'm describing above?

I actually run jumbo (9000) on my local LANs and the tunnel is only 1480.

There really is no problem with this and there is no slow down of any
kind, indeed there is a initial PTB, but who cares about that, it will
happen anyway as the server side will typically be 1500.

I guess you see the difference quite well there that way as local
transfers are 9000 while internet traffic is 1480, but then again
currently LAN is GigE while internet is only 25/3 which is another major
difference.


I am really wondering why people would advise misconfiguring the MTU on
a local network. It really is not needed.


One note I have to make that it might be useful, but a bit late to have
such a change, if there was a prefix inside the ICMP PTB. That way a
gateway sending a PTB could send a prefix along with the information,
and if the recipient accepted that information as trusted it could limit
everything in that destination to 1280 for instance. There are lots of
problems with the 'trust' portion there though, thus likely the only
'trust' would be if that went in sizes of /64s...


On 2013-01-15 06:52, Mikael Abrahamsson wrote:> On Mon, 14 Jan 2013,
Jeroen Massar wrote:
>
>> The only thing that having 1480 in the RA might do is that some OSs
>> might advertise 1480 in the TCP MSS which does get to servers who are
>> stupid (really no other way to say that ;) and filter ICMP PTB and
>> then the problem is partially circumvented.
>
> "some OSs"? Which one wouldn't?

I am not 100% about the statement (hence the might), but I would assume
some smart OS might be doing this. I would if I wrote a TCP stack ;)

> Why "filter PTB"? With a lowered MSS (derived from the lower IPv6 MTU
> seen in the RAs) there will never be any PTBs needed.

There are people who use IP for other things than TCP..... this little
UDP thing also still exists next to SCTP and various others.

If PTB (or even full ICMP) was not broken/filtered in the first place
things just works, they do for me...


Also, against setting the MTU to <1500 on the LAN, what if an
far-away-upstream does a tunnel of 1280, does that mean "oh lets just go
1280 everywhere" ? :)

Greets,
 Jeroen