Re: IETF Last Call conclusion for draft-ietf-6man-rfc2460bis-08

[Tim Chown <[e-mail]>]

Hi,

> On 30 Mar 2017, at 19:19, 神明達哉 <[e-mail]> wrote:
> 
> At Thu, 30 Mar 2017 11:52:41 -0500,
> Robert Raszuk <[e-mail]> wrote:
> 
>> Ok so till a new document updates 2460bis any further work on EHs is frozen
>> as it would reference 2460bis with new text. That was my main point.
> 
> I don't get the logic...an "update" can start immediately once such a
> draft new proposal is available.  The update won't be formally
> completed until it gets some formal state like a standard track RFC,
> and it will take time, but that wouldn't necessarily mean a further
> work is "frozen"; it's not very clear to me what this term means in
> this context, but it's quite common development takes place while the
> spec is being discussed as a draft, and it's also not uncommon some
> commercial operators even start deploying it.  On the other hand, even
> if we now agreed that rfc2460bis should explicitly allow such "further
> work", the discussion itself would take long and wouldn't be completed
> soon.
> 
> But IMO it's irresponsible to leave the text ambiguous and let some
> other people misunderstand it, possibly even more casually and/or in
> the global Internet, for the comfort of some particular future work.
> I think we're now trying to help avoid the latest clarification in
> rfc2460bis to be interpreted as an "outright ban" of future updates
> while still trying to be responsible for the soundness of the global
> Internet.  In my understanding Brian's additional text is one such
> attempt (I also proposed text in that sense at the time of WGLC,
> although it wasn't adopted in the end).  If that text is still not
> enough we can discuss how to phrase it.  And, while I suspect people
> who wanted to keep the ambiguity will never be satisfied with the
> result as long as the added clarification remains, I believe that's a
> reasonable compromise to achieve a balance between being responsible
> and not (unintentionally) discouraging future updates.

Repeating what I said at the mic, my views echo those of Jinmei, Michael, Lorenzo and Brian. I agree with the decision to progress RFC2460-bis to IS including Suresh’s text. I’d also like to see Brian’s short amendment included, which makes explicit Internet scope as the default target.

Advancing 2460bis doesn’t freeze work on SR; that work can continue today. I don’t see the new SR header insertion draft completing in “a couple of weeks”; it’s much more likely to take several months or even a year to progress through WG adoption and eventual publication, to ensure good and robust input. For adoption, I’d expect many people to want the draft to include text on a) why vanilla encapsulation is not sufficient and b) how the insertion can guaranteed to be done safely for specific scenario(s). But that’s all very possible.

And please do note that publishing 2460-bis as an Internet Standard does not set it in stone. Your SR work can update it, as could other future RFC(s) on other new mechanism(s) or innovation(s) that have yet to be proposed. I’d argue that it’s pretty likely we’ll see other RFCs update 2460-bis; that shouldn’t be a surprise if it’s a protocol that we’ll be using for the next 20-30 years or more. For each, we’ll just see “Updated by:” tags in the RFC header. The current RFC2460 has accumulated nine such tags in 19 years, so about one every two years.

On enterprise deployment I’d argue that, if anything, moving to IS should help strengthen the case for deployment, because it confirms proven widescale deployment.  The four criteria for moving to IS are described in Section 2.2 of RFC6410, but are summarised as "a high degree of technical maturity and by a generally held belief that the specified protocol or service provides significant benefit to the Internet community”. I’d hope we have very strong consensus on that statement.

Tim