Re: [trill] Charter Revision TRILL Link Types Item

[Sam Aldrin <[e-mail]>]

Hi Donald,

Inline for my comments.

Sent from my iPhone

On May 16, 2013, at 5:09 AM, Donald Eastlake <[e-mail]> wrote:

> Hi Sam,
> 
> On Tue, May 7, 2013 at 11:02 AM, Sam Aldrin <[e-mail]> wrote:
>> I am not comfortable with new text either.
>> It should rather say, TRILL extensions to support extension of TRILL campus over WAN/internet. Something like that.
> 
> I see two parts to what you want: (1) various link technology
> protocols, and
Yes
> (2) optionally a more restricted control protocol to
> reduce the probability of error propagation, etc.
No. I didn't imply that. See below for details.
> 
> Looking at these separately:
> 
> (1) Generally TRILL over X link technology stuff gets done in the X
> WG, if there is an X WG, not in the TRILL WG. For pseudo wires, we
> have agreement from the PWE3 Chairs and AD that it should be done in
> TRILL as long as there is not change to PWE3. For IP, there is no IP
> WG. I'm not convinced we could get something that would imply we could
> do TRILL over an WAN technology approved in our Charter.
IP is unique as there in no WG as such. Understand that. The premise of my comment is as follows. We do anticipate changes in TRILL like adding new TLV or something else, while extending TRILL over IP or PWE3 or other types. Instead of doing per WAN/link  type, considering general extensions for TRILL such that, one could extend TRILL over IP or MPLS or VXLAN or whatever. That is better effort spent by the WG rather than doing extensions again and again per each type. If this produces multiple drafts or not in not at all the point, but charter item shouldn't restrict to one or two types.
> 
> (2) You might want to run the currently specified TRILL protocol over
> a WAN link or you might want to run a reduced control protocol. This
> is mostly orthogonal with the link technology. We already have the
> current TRILL protocol and doing a reduced protocol is a separate item
> in the draft Charter.
See my above answer why I didn't imply that. If protocol doesn't already do that, (which it does as you say), we have bigger problem to deal first.
> 
> So, for the above reasons, I think the draft Charter item is reasonable.

I do not think it is.
> 
>> Narrowing charter it to fit specific drafts or solutions will keep other work, outside of it, not in scope.
> 
> Drafts mentioned in the draft Charter are just supposed to be
> examples, not binding, and could be dropped.
If they are not binding, drop them.

Sam
> 
> Thanks,
> Donald
> =============================
> Donald E. Eastlake 3rd   [phone] (cell)
> 155 Beaver Street, Milford, MA 01757 USA
> [e-mail]
> 
>> -sam
>> 
>> On May 2, 2013, at 7:06 PM, Donald Eastlake <[e-mail]> wrote:
>> 
>>> Hi,
>>> 
>>> "Item (4) Support of IP links between TRILL switches, for example to
>>> connect branch office TRILL switches to a central campus over the
>>> Internet.  (draft-mrw-trill-over-ip)"
>>> 
>>> Based on the discussion at the meeting, including the discussion of
>>> draft-yong-pwe3-trill-o-pw, which has been tweaked and re-posted as
>>> draft-yong-trill-o-pw, I suggest this item be replaced by
>>> 
>>> "Item (4) Support of TRILL over pseudowires and IP tunnels between
>>> TRILL switches, for example to connect branch office TRILL switches to
>>> a central campus over the Internet. (draft-yong-trill-o-pw)"
>>> 
>>> Comments?
>>> 
>>> Thanks,
>>> Donald
>>> =============================
>>> Donald E. Eastlake 3rd   [phone] (cell)
>>> 155 Beaver Street, Milford, MA 01757 USA
>>> [e-mail]
>>> _______________________________________________
>>> trill mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/trill
>>