Minutes of:
Mobility for IPv6 (mip6) WG
At: IETF69 (Chicago)
TUESDAY, July 24, 2007 from 1740-1840 

Chairs: Basavaraj Patil <basavaraj.patil@nsn.com> 
	Alper Yegin <alper.yegin@yegin.org)

Credit for these minutes:

1. Basavaraj Patil

2. Wassim Haddad (wmhaddad@gmail.com)

====================================================

- Status of documents presented by Alper
  Refer to slides

- Yoshi presented an update on the protocol document
  Refer to slides

Issue on Version field:
- Alper said that removing the version number is okay. Had a
  discussion with Sam Hartman who is okay with the proposal

Lionel: 

Alper: Are you trying to find a solution for capability negotiation?

Lionel: Thinks that version field can be in an AVP

Alper: Is possible to do if we had an error indicating mechanism. For
dealing with version numbering, it is not critical to add error
handling. 

Jari: Not sure if the Diameter model should be considered in this case
as it is different. For protocols that use AVPs, you need to consider
different mechanisms. The current document is okay. Worried that there
are no error messages. But this concern is agnostic of the version
number one. 

Lionel: Is there no version number in the messages? How does PANA
indicate that the protocol does not support some AVP?

Yoshi: Extensions can be developed to deal with the problem. 

Alper: Two mechanisms exist: Either we have a capability negotiation
mechanism or dynamically figure it out

Lionel: The issue is not for new AVPs. It is for

Mark: Version number in the header is a bad idea and so getting rid of
it is good. If the version remained it would only indicate the format
and the stacking of AVPs. 
If a mandatory AVP is not supported, terminate the session and provide
an appropriate result code. 

Yoshi: PANA terminate message is only set during the
authentication/authorization phase and not at any time.

Mark: Is there a good reason why this is the case?

Alper: Current spec is sufficient in terms of how the protocol handles
mandatory AVPs. 

Jari: Do you really need the "m" bit?
Mark: It is really not all that great.

Conclusion: Get rid of the version number, No mandatory field in
AVPs. All future AVPs would have to have a negotiation capability. 

---------------------------

Issue on Security considerations:

Alper is resistant to mandating any securiy on channel bindings. He
will propose some text for discussion with Sam.