Re: IPV6 Doubt

Carlos, after reading both the Cisco paper and the one written by Ivan that
you referenced, it seems that UDP traffic has to be handled as an
exception. That would explain at least to me why the Cisco paper says it is
TCP only feature.

From Ivans paper: "The TCP stack in the originating host or a PMTUD-aware
UDP application has to retransmit the data in smaller datagrams." The PMTUD
function seems to be inherent in TCP but requires special treatment for
UDP. Thats what I get out of it anyway.

On Tue, Jun 7, 2011 at 7:51 PM, Carlos G Mendioroz <tron_at_huapi.ba.ar> wrote:

> AFAIK, PMTUD has no need for TCP.
> And is not directly related to MSS tinkering, other than that being a
> band-aid fix for the overzealous security admins filtering all ICMP
> messages and breaking PMTUD.
>
> You may read http://www.nil.com/ipcorner/IP_Fragmentation/ by Ivan
> Pepelnjak for a nice writeup on the subject.
>
> -Carlos
>
> Joe Astorino @ 07/06/2011 20:29 -0300 dixit:
>
> My thought originally was it may have to do with the idea that TCP can
>> retransmit segments while UDP does not but I could be wrong. Maybe someone
>> else knows definitively.
>>
>>
>> Sent from my Verizon Wireless BlackBerry
>>
>> Regards,
>>
>> Joe Astorino
>> CCIE #24347
>>
>> "He not busy being born is busy dying" - Dylan
>>
>> -----Original Message-----
>> From: Carlos G Mendioroz <tron_at_huapi.ba.ar>
>> Date: Tue, 07 Jun 2011 20:23:46 To: <joeastorino1982_at_gmail.com>
>> Cc: Routing Freak<routingfreak_at_gmail.com>; Cisco certification<
>> ccielab_at_groupstudy.com>
>> Subject: Re: IPV6 Doubt
>>
>> But you said it made sense... :)
>>
>> Joe Astorino @ 07/06/2011 20:15 -0300 dixit:
>>
>>> That is what it states in the document I posted. Check it out
>>>
>>>
>>> Sent from my Verizon Wireless BlackBerry
>>>
>>> Regards,
>>>
>>> Joe Astorino
>>> CCIE #24347
>>>
>>> "He not busy being born is busy dying" - Dylan
>>>
>>> -----Original Message-----
>>> From: Carlos G Mendioroz <tron_at_huapi.ba.ar>
>>> Date: Tue, 07 Jun 2011 20:06:52 To: Joe Astorino<
>>> joeastorino1982_at_gmail.com>
>>> Cc: Routing Freak<routingfreak_at_gmail.com>; Cisco certification<
>>> ccielab_at_groupstudy.com>
>>> Subject: Re: IPV6 Doubt
>>>
>>> Joe,
>>> why do you think that PMTUD is only possible under TCP ?
>>> All it takes is sending oversized packets with DF, and waiting for
>>> someone to tell back "no go". Any IP packet can do that, or am I missing
>>> something ?
>>> -Carlos
>>>
>>> Joe Astorino @ 07/06/2011 12:29 -0300 dixit:
>>>
>>>> From reading this document it seems that path MTU discovery (PMTUD) is a
>>>> way
>>>> for routers to determine the maximum IP MTU they can use along a path.
>>>> Seems it is only possible for TCP traffic according to the doc which
>>>> makes
>>>> sense because if you send a TCP segment and it is higher than the
>>>> outgoing
>>>> MTU at some point along the path, you should get an ICMP message back
>>>> saying
>>>> so and be able to retransmit that segment. "ip tcp path-mtu-discovery"
>>>> seems to be the same basic concept but for router initiated traffic.
>>>>
>>>> Now, for IPv6 I believe that the PMTUD function is moved off of the
>>>> routers
>>>> and on to the hosts but basically that is how I read things here.
>>>>
>>>>
>>>> http://www.cisco.com/en/US/tech/tk827/tk369/technologies_white_paper09186a00800d6979.shtml#t4
>>>>
>>>> HTH
>>>>
>>>>
>>>> On Tue, Jun 7, 2011 at 11:04 AM, Routing Freak <routingfreak_at_gmail.com
>>>> >wrote:
>>>>
>>>> Hi fellows,
>>>>> I am searching for a command to enable path mtu discovery in IPV6.
>>>>> I want to know the difeerence between ip mtu discovery and ip tcp path
>>>>> mtu
>>>>> discovery
>>>>>
>>>>> Can some one point out the difference.
>>>>>
>>>>>
>>>>> Blogs and organic groups at http://www.ccie.net
>>>>>
>>>>> _______________________________________________________________________
>>>>> Subscription information may be found at:
>>>>> http://www.groupstudy.com/list/CCIELab.html
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>
> --
> Carlos G Mendioroz <tron_at_huapi.ba.ar> LW7 EQI Argentina
>

-- 
Regards,
Joe Astorino
CCIE #24347
Blog: http://astorinonetworks.com
"He not busy being born is busy dying" - Dylan
Blogs and organic groups at http://www.ccie.net