Re: [trill] Should we make draft-mrw-trill-over-ip-01 a WG document?

[Joe Touch <[e-mail]>]

Regarding this approach, I'm also wondering why TRILL-over-IP?

Since TRILL is over ethernet, and ethernet is over IP, why is a specific 
solution for TRILL over IP even needed? Seems like a lot of potential 
complexity to save 18 bytes.

Joe

On 5/14/2012 8:04 AM, Tissa Senevirathne (tsenevir) wrote:
> Hi James
>
> I think we are agreeing on lots points except one important point i.e.
> SHOULD or SHOULD NOT make individual submissions WG docs without a
> strategy.
>
> My main objection of making individual submissions WG docs without well
> thought through strategy is it takes away the focus and confuse WG,
> customers and burden vendors. Additionally, TRILL WG should focus on the
> priority items that was agreed in Taipei. Adding additional WG docs is
> not helping anyways. I am in the opinion that WG should identify
> priority Work areas, then form strategy around it, bring together
> different ideas to form well thought through solutions.
>
> Secondly, it seems you are in the opinion that to get WG input the draft
> need to be WG document. I disagree with that. A draft does not need to
> be in WG status to get WG feedback. Authors should solicit feedback
> through the mailing list. Making a WG status is not going to change
> that.
>
>
>
> -----Original Message-----
> From: James Carlson [mailto:[e-mail]]
> Sent: Monday, May 14, 2012 7:30 AM
> To: Tissa Senevirathne (tsenevir)
> Cc: Erik Nordmark; [e-mail]
> Subject: Re: [trill] Should we make draft-mrw-trill-over-ip-01 a WG
> document?
>
> Tissa Senevirathne (tsenevir) wrote:
>> That sounds a bit backwards to me.
>>
>> [Answer] Doing small piecemeal work, is not the right approach.
>
> We might agree on that point, but it's not the question at hand.  The
> question at hand is whether to adopt a draft.
>
>> The document is clearly in scope of the working group, and I believe
>> the only other question regarding adoption is whether the wg wants
>> (and the authors want) to steer the document by the working group
>> consensus process.
>>
>> [Answer] Good point, I re-read the charter one more time and there is
>> no where it says TRILL over IP or other encapsulation, so it seems the
>
>> document is not in charter. Please could you point to the WG, which
>> line item in charter that qualifies this work to be in scope ?
>>
>> https://datatracker.ietf.org/wg/trill/charter/
>
> Sure:
>
> 	"This includes a MIB module and
> other pieces needed for operations, but also additional ways to extend
> and optimize TRILL for the properties of the networks on which it is
> deployed."
>
> This seems to me to be an additional way to extend TRILL for the
> property of a network -- specifically, an IP network.
>
> Granted, I'm not thrilled about the idea of adding a new means to reach
> mutual encapsulation.  I assume the authors have a good reason to do
> this, and have interesting ideas on how and why it should be done.  The
> important point is that I think this working group is the best possible
> forum to air those ideas.
>
>> [Answer] IETF tradition has been to start with a problem space and
>> create a common solution not making every draft that is in scope of
>> charter a WG document. Your point on "too many solutions" exactly that
>
>> is my point having subset of solutions finally leads to too many small
>
>> pieces.
>
> Some efforts do indeed follow that simple linear path -- idea, BOF,
> charter, WG, publication.  But that's certainly not the only path, nor
> perhaps even the most common.  Adoption by an existing WG of new
> in-scope documents is something that I view as crucially important to
> continued WG viability.  Without it, protocols just grow like weeds as
> new extensions are otherwise developed, deployed, and documented without
> the help of WG consensus using the individual-submission RFC publication
> track.
>
> Failing to adopt does not foreclose publication of an RFC describing
> this extension, nor any other work by the authors.  It simply
> disconnects the WG from the process.
>
> Sending TRILL over IP off to the individual-submission world without WG
> input would, I think, be a poor result of this exercise.  I've been
> through that problem as PPPEXT chair, and it's no fun.
>