Re: [Isis-wg] New Layer2 extensions to ISIS - thorough review will be required

["Fedyk, Donald (Don)" <[e-mail]>]

I agree with Dave Allan.

This work did not start overnight for 802.1aq and TRILL. 
We were asked to put 802.1 and TRILL together and it turns out we are orthogonal. While there is a little bit of sharing of top level TLVs our content is different. The value of having them in one draft is we can see this, however to date I doubt it has made the draft any more readable. It is sad to say we only realized that OTV was in there when comparing notes in draft 3. The only thing the technologies share is the desire to carry MAC addresses. To review the draft it must be readable so I don't see it as simply a nit. 

>From and 802.1aq standpoint we are happy with no new PDUs.  IS-IS carries multiple address families to date, a capability that we wanted to leverage. 

Regards,
Don 


-----Original Message-----
From: [e-mail] [mailto:[e-mail]] On Behalf Of David Allan I
Sent: Saturday, April 24, 2010 12:31 PM
To: David Ward; isis-wg
Cc: Christian Hopps; Adrian Farrel; Ralph Droms
Subject: Re: [Isis-wg] New Layer2 extensions to ISIS - thorough review will be required

Hi Folks:

While the decision of splitting the drafts is in discussion (and that is a direction I favor as the scope of IS-IS changes is radically different and disjoint between the "layer 2"s), the question needs to be asked "what in the draft survives the split as WG items". 

802.1aq - yes
TRILL - yes
OTV - does not belong here, it is proprietary... 

If OTV is standardized it COULD be considered by the WG at some point in the future. Right now it is simply an wholly inappropriate complication in addressing the two technologies the WG has agreed to...

It should be removed from the draft and a new version published ASAP.

My 2 cents
D

-----Original Message-----
From: [e-mail] [mailto:[e-mail]] On Behalf Of David Ward
Sent: Friday, April 23, 2010 1:30 PM
To: isis-wg
Cc: Christian Hopps; Ralph Droms; Jari Arkko; Adrian Farrel
Subject: [Isis-wg] New Layer2 extensions to ISIS - thorough review will be required

All -

The WG Chairs would like to be clear on the approach to review and progression of the draft: draft-ietf-isis-layer2. There will be a request for the routing directorate, specific guest reviewers and a lengthy community review. There will also be a requirement of multiple, interoperable implementations. Whether we split the draft into TRILL, .aq, OTV extensions into separate drafts or not, the foundational changes to ISIS for each of the technologies will need to be proven working and correct.

The draft itself will need to be cleaned up and clear what PDUs, TLVs, etc are necessary for any of this technology which will make reading and implementation awareness easier. If we decide to split them (we will decide later - please remain calm  and not discuss this nit now- let's get the information completed as we know it); it will be quite trivial. If we decide not to; it will be clear to the reader/implementor.

The purpose of such a process is because in several decades of the existence of ISIS and extremely wide deployment, no new PDU types were ever deemed necessary. The new TLVs being proposed and the data being carried has never been carried before and may cause new scaling, performance or other impacts that we don't currently understand. We need to fully understand the impact of these changes and learn what we don't know about the changes we are proposing. I've found in my experience that new protocols, new technologies and fundamental changes to protocols require a deep understanding and implementation experience and not a rush to RFC. Headsup

Thanks.

-DWard, CHopps

_______________________________________________
Isis-wg mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/isis-wg
_______________________________________________
Isis-wg mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/isis-wg