Re: Communication in response to the OIF

[Stephen J Trowbridge <sjtrowbridge@lucent.com>]

Hi Fred,
I think rather than any kind of knee jerk reaction, it is probably better to
think before we act.

It isn't really important whether we like the situation as it is now and whether
or not we like what OIF has done. What we should consider is, starting from
where things stand now, what sort of response is likely to lead to the best
result for IETF and for the industry?

I think it is a good bet that the OIF demo is going to occur whether or not we
like what they have done. Participants have made investments to make this happen
and it isn't going to stop now. And as you say, even with the "experimental"
characterization of what they are doing, a demonstration like this in such a
high profile venue will surely develop a measure of credibility for the solution.

If we take this as a given, the next questions I would ask are:
- Is it better for IETF to disavow any association with the OIF effort, with the
result being that multiple "descendants" of the GMPLS protocols (whether OIF is
allowed to use the "GMPLS" term or not) compete with each other in the market;
or would it be better to try to constructively engage with OIF to try to limit
the proliferation of solutions and to bring the OIF solution in line with IETF
principles?

- Is it better in the demo if IETF gets credit (and presumably good press) for
the base protocols that were employed to make it happen, or should this look
like OIF did this single-handedly? (I would guess that ITU-T will probably ask
to get credit for the technology from their organization that was used to make
the demo happen. Shouldn't we do the same?).

Regards,
Steve

On 3/9/2004 8:28 AM, Fred Stringer wrote:
> Hi Kireeti,
> A comment from the lurking gallery if I may.
> You are probably representing the committee opinion here, but there seems 
> to be a conflict in the  message.
> You stated that since the work was emphasized by Mr. Ong as "experimental" 
> alleviated some concerns - but then there is the concern over the SuperComm 
> testing.
> I don't think INTEROPERABILITY demonstration in public forum is purely 
> experimental.
> I don't see why the concerns are alleviated. You don't want to clobber 
> Lyndon but I think the concern is justified. The industry has enough 
> problems without fracturing the protocols it depends upon.
> 
> The request of not using OSPF, OSPF-TE and GMPLS in the demo has teeth and 
> is good.
> 
> Of all the comments you were probably expecting I'm sure this is not one of 
> them. But the situation did move me to come a little out of my lurking status.
> 
> cheers
> Fred
> 
> 
> At Tuesday09:13 AM 3/9/2004, Kireeti Kompella wrote:
> 
>>Hi All,
>>
>>Here's a first draft of a reply to the OIF.  Please comment to the
>>list by Monday, March 15 2004.
>>
>>If someone has email addresses for Steve Joiner, Jim Jones and John
>>McDonough (and titles for the last two), that would be very helpful.
>>
>>Thanks,
>>Kireeti.
>>--------
>>
>><Date>
>>
>> From: Kireeti Kompella & Adrian Farrel, IETF CCAMP Working Group Chairs
>>
>>To: Mr. Steve Joiner, OIF Technical Committee Chair
>>Cc: Jim Jones,
>>Cc: John McDonough,
>>Cc: Alex Zinin,       IETF Routing Area Director
>>Cc: Bill Fenner,      IETF Routing Area Director
>>
>>Dear Steve,
>>
>>Thank you for your communication regarding the current status of OIF
>>signaling and routing work, and the associated documentation.  This
>>communication is in response.  As there is no formal liaison
>>relationship yet between the IETF and the OIF, this communication is
>>not treated as a liaison; hopefully, this situation will be rectified
>>soon.
>>
>>Thank you too for allowing Mr. Lyndon Ong to present a synopsis of
>>the work going on at the OIF with regard to Intra-carrier E-NNI
>>routing.  It was both useful and enlightening.
>>
>>However, both of these gave us cause for alarm, on two fronts:
>>a) The development of new or modified code points and procedures
>>   in OSPF without expert review from the OSPF WG in the IETF
>>   contravenes IETF procedure, especially as the IETF pays special
>>   attention to changes in protocols in the Routing Area, such as
>>   OSPF.
>>b) The development of routing for optical networks without expert
>>   review from the CCAMP WG is also a source of concern, especially
>>   in the light of a cooperative effort between the ITU-T and the
>>   IETF in exactly this area.
>>
>>Mr. Ong's emphasis that this work was experimental and purely for the
>>purpose of testing alleviated some of our concerns.  It would be very
>>helpful to hear this officially from the OIF; furthermore, in the
>>interests of openness and full disclosure, we would strongly urge the
>>modification of a paragraph in the Introduction of the draft routing
>>document OIF2003.259 as follows:
>>
>>   "The base protocol as defined by this document is OSPF with
>>    extensions for Traffic Engineering and GMPLS.  This document
>>    proposes to use GMPLS-OSPF to operate at each hierarchical
>>    level, with multiple such levels stacking up to form the
>>    routing hierarchy.  Extensions have been defined in the forms
>>    of (sub-) TLVs to accommodate the requirements as defined in the
>>    G.8080, G.7715, and G.7715.1.  Note that these extensions as
>>    currently specified are purely for the purpose of experimentation
>>    and testing; in particular, they have not yet been reviewed by
>>    the OSPF and CCAMP Working Groups in the IETF.  Furthermore they
>>    use experimental codepoints, and as such must not be used in
>>    production deployments."
>>
>>Mr. Ong also brought to our attention that the OIF will be holding
>>an interoperability demonstration of this specification at the
>>SuperComm in June 2004.  Due to the preliminary nature of this
>>specification, the IETF would strongly recommend that the words
>>OSPF, OSPF-TE and GMPLS not be used in the context of this
>>demonstration, nor that there be any implication that this work
>>has been reviewed or sanctioned by the IETF.
>>
>>It would be helpful in determining the future relationship between
>>the IETF and the OIF to understand how the OIF intends to progress
>>this document.
>>
>> o Is this intended to become an Implementation Agreement in
>>   something close to its current form?
>>
>> o Does the OIF intend to submit this as a submission in the ITU-T
>>   SG15 to become a Recommendation?
>>
>> o Does the OIF intend to submit this document as an Internet Draft
>>   to become an IETF RFC?
>>
>>Thank you for your attention in this matter, and for initiating this
>>dialogue.  We hope that this develops into a fruitful relationship.
>>To that end, we enclose a product of the joint work between the
>>ITU-T and the IETF on Routing Requirements for ASON.  This is a
>>work in progress, and we solicit your comments:
>> - to identify any requirements that the OIF has over and above those
>>    requirements established by the ITU-T ASON model
>> - to ensure that the solution developed within the IETF addresses
>>    the requirements of both the ITU-T and OIF.
>>
>>We hope that your feedback will signal the beginning of an active
>>cooperation between the IETF and the OIF.
>>
>>Sincerely,
>><etc.>
>>
>><attachment: current version of GMPLS ASON Routing Requirements doc>
> 
> 
> 
>