Re: [v6ops] Spencer Dawkins' No Objection on draft-ietf-v6ops-pmtud-ecmp-problem-04: (with COMMENT)

[Spencer Dawkins at IETF <[e-mail]>]

Hi, Benoit,

On Thu, Oct 15, 2015 at 2:47 AM, Benoit Claise <[e-mail]> wrote:

> Spencer,
>
> Spencer Dawkins has entered the following ballot position for
>> draft-ietf-v6ops-pmtud-ecmp-problem-04: No Objection
>>
>> When responding, please keep the subject line intact and reply to all
>> email addresses included in the To and CC lines. (Feel free to cut this
>> introductory paragraph, however.)
>>
>>
>> Please refer to https://www.ietf.org/iesg/statement/discuss-criteria.html
>> for more information about IESG DISCUSS and COMMENT positions.
>>
>>
>> The document, along with other ballot positions, can be found here:
>> https://datatracker.ietf.org/doc/draft-ietf-v6ops-pmtud-ecmp-problem/
>>
>>
>>
>> ----------------------------------------------------------------------
>> COMMENT:
>> ----------------------------------------------------------------------
>>
>> In section 3, "Mitigation", I'm reading this.
>>
>>     Mitigation of the potential for PTB messages to be mis-delivered
>>     involves ensuring that an ICMPv6 error message is distributed to the
>>     same anycast server responsible for the flow for which the error is
>>     generated.  Ideally, mitigation could be done by the mechanism hosts
>>                 ^^^^^^^
>>     use to identify the flow, by looking into the payload of the ICMPv6
>>     message (to determine which TCP flow it was associated with) before
>>     making a forwarding decision.  Because the encapsulated IP header
>>     occurs at a fixed offset in the ICMP message it is not outside the
>>     realm of possibility that routers with sufficient header processing
>>     capability could parse that far into the payload.  Employing a
>>     mediation device that handles the parsing and distribution of PTB
>>     messages after policy routing or on each load-balancer/server is a
>>     possibility.
>>     Most of the document is about using other mitigations that make this
>> mitigation unnecessary, but it's still a bit odd to see Deep Packet
>> Inspection characterised as "ideally", now that RFC 7258 is now BCP 188,
>> and we've chartered TCPINC to make that DPI fail as often as possible.
>>
> From a resource consumption point of view, the ideal situation would be
> NOT to distribute the PTB message to all anycast servers. This is a fact.
> Therefore, I don't have a problem with "ideally" in this context. After
> all, this document is not a spec on using DPI.


I agree with your observation. I would think that in order for something to
be ideal, it should be possible (ideally we should be able to send packets
at greater than the speed of light, but we can't, so. Ideally we should be
able to send the PTB only to the place it needs to go, but we can't, so).

But I'm a TSV guy who is poking at this because of RFC 7258, so if the SEC
folk don't care, I'm fine.

Spencer


> Regards, Benoit
>
>>
>> Could that one word go away?
>>
>> I see in email that Fred asked if any reviews were needed before
>> submitting this draft for publication, and David Black suggested reviews
>> from INT and TSV, and that made it as far as the shepherd writeup, but
>> I'm not seeing a request for review popping up in TSV (at least not with
>> "PMTUD" and "V6OPS" in an e-mail). I don't have concerns about this
>> particular draft, but I wish it hadn't slipped through that particular
>> crack.
>>
>> Yes, Fred copied TSVWG and TSVAREA on his note asking for guidance, and
>> yes, there was an IETF Last Call, so I'm not hitting "Defer" to chase
>> that. Just something to watch for, in the future.
>>
>> And yes, Martin and I have a call set up to talk with the TSVdir triage
>> team next week ...
>>
>>
>> .
>>
>>
>