[trill] draft-ietf-trill-arp-optimization WG Last Call (11/15

[trill] draft-ietf-trill-arp-optimization WG Last Call (11/15 - 12/6)

"Susan Hares" <shares@ndzh.com> Sat, 14 January 2017 00:43 UTC

From: Susan Hares <shares@ndzh.com>
To: trill@ietf.org
Date: Fri, 13 Jan 2017 19:39:44 -0500
Message-ID: <00dc01d26dfe$b3f36800$1bda3800$@ndzh.com>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----=_NextPart_000_00DD_01D26DD4.CB1EE6A0"
Thread-Index: AdJt/qEeRDXjbkFHSnaNmsUQMopQxA==
Content-Language: en-us
Archived-At: <https://mailarchive.ietf.org/arch/msg/trill/xMxINbUDrWCMl8nh0cFrcP0arB8>
Cc: 'Donald Eastlake' <d3e3e3@gmail.com>, akatlas@gmail.com
Subject: [trill] draft-ietf-trill-arp-optimization WG Last Call (11/15 - 12/6)
Precedence: list

Re-forwarding to get the right link to the WG LC.

Sue 

From: Susan Hares [mailto:shares@ndzh.com] 
Sent: Saturday, January 7, 2017 8:59 PM
To: 'Donald Eastlake'; 'Liyizhou'; linda.dunbar@huawei.com; 'Radia Perlman';
'Mohammed Umair'
Cc: 'Alia Atlas'
Subject: Donald Eastlake 

Donald, Yizhou, Linda, Radia, and Mohammed: 

I reviewed the TRILL ARP/ND optimization this evening (after my head finally
clear from allergies).  I do not believe you have answered Suresh Krishnans
issues about SEND clearly.   I see 80% of the logic, but no place where you
clearly indicate when SEND should be used.   I think you have answered 90%
of Alvaros concerns, but Im concerned that without specific references to
SEND and clear indication about the SEND messages. 

Heres a set of changes that I think might work to resolve.     

Cheerily, 

Sue 

----------

#1 

Page 5  section 4, title

Old:/ Handling ARP/ND messages/

New:/Handling ARP/ND/SEND messages/

#2  change paragraph #3  to add SEND 

                /in ND/ to  /in ND and SEND/

#3  Could you as section 4.1,  a short description about send that
includes: 

1)      Why Send is needed, 

2)      How send interacts with ND, 

3)      Intermixing Topology  is this topology allowed

SEND ßà RBRIDGE ------RBRDGE ßàND 

[IMHO  this is a.2 or could be part of a.5] 

4)      When SEND should be used 

#3  Could you add a section 4.2  1 paragraph that indicates why or why not
the RBRIDGE would verify the address. 

#4  section 4.1 (change to section 4.3) 

                Add send by OLD%ARP/ND% to New%ARP/ND/SEND% 

#5  section 4.2 (change to section 4.4) 

5.1 Paragraph 2, item a) 

change ARP/ND request to  ARP/ND/SEND request 

Change a.2 from ARP/ND to ARP/ND/SEND 

Change a.5 from ARP/ND to ARP/ND/Send request 

5.2 paragraph 3  item b)  

Change sentence 1:   Old% ARP/ND%  New% ARP/ND/SEND%

Change sentence 3 end 

Old:/targets private key and only action b.1/

New/targets private key therefore only action b.1/ 

5.2 paragraph 4  item c)

Change sentence 1:  OLD%ARP/ND% NEW%ARP/ND/SEND%

5.3 paragraph 5) item d) 

Include in the DAD [RFC4862] the ND and SEND version 

#6  describe what happens if SEND CGA option does not exist in send, or
fails

This can be a new section or a re-written section 7 or a sub-section to
section 7.

Describe what happens SEND CGA exists, does not exist, or fails due to the
wrong key.   

This could also go in a revised section 4.1 

Section 8  expand this to include the SEND functionality.  

Thanks

[trill] draft-ietf-trill-arp-optimization WG Last… Susan Hares