DAD already has rough consensus and running code?  What are you talking
about?

This is a bit more than a research issue now IPv6 products are shipping.

/jim

> -----Original Message-----
> From: Francis Dupont [mailto:[EMAIL PROTECTED]]
> Sent: Monday, March 04, 2002 4:28 AM
> To: Margaret Wasserman
> Cc: Hesham Soliman (ERA); [EMAIL PROTECTED]
> Subject: Re: Should DAD be optional? [Was
> draft-ietf-ipv6-cellular-host-00.txt -> wg last call?] 
> 
> 
>  In your previous mail you wrote:
> 
>    >   > It is not my impression that the WG has reached 
> consensus on some
>    >   > of the issues raised in this document, specifically:
>    >   > 
>    >   >          - Forbidding the use of DAD on some links
>    >
>    >=> DAD is not needed if address duplication is not
>    >possible. Since each terminal gets a unique
>    >/64 (as per the advice draft), and since all terminals
>    >are on p2p links, DAD would be a waste of BW.
>    >But please have a closer look and let us know 
>    >if something was missed. 
>    
>    Address duplication is possible on point-to-point links, because
>    one end may choose an address that is already in use by the other
>    end.  I don't think that this is very likely, but it is possible.
>    
> => I disagree for addresses using the explicitely negociated
> interface IDs with in the first position the link-local addresses.
> 
>    However, making DAD optional (and advising against the use of DAD) 
>    on point-to-point links, is in direct conflict with RFC 2462,
>    which says:
>    
>    "Duplicate Address Detection MUST take place on all unicast 
>    addresses, regardless of whether they are obtained through 
>    stateful, stateless or manual configuration..."
>    
> => the point-to-point link is one case DAD is overkilling.
> But as too many persons want to make DAD optional I agree
> we have to be carefull. For instance we should ask that
> all DAD stuff must get rough consensus in the IPv6 WG.
> Another point is who makes the choice to avoid the DAD check:
> my opinion is this must be the "link manager" (i.e. the manager
> of the network where is the link).
> 
>    I don't think that we should publish an informational 
> document that 
>    advises some implementors to do something that specifically
>    disagrees with a MUST requirement in a standards-track document.
>    If the standards-track document is broken, we need to fix it 
>    instead.
>    
> => I agree we have a procedural case at least.
> 
>    [Please note that I actually think that we should be able to 
>    disable DAD for some link types.  I made a proposal to that effect
>    several years ago, but my arguments didn't win the day.]
>    
> => so basically we agree!
> 
> Thanks
> 
> [EMAIL PROTECTED]
> 
> PS: DAD is not a toy, at least one time I saw it saved us 
> from a disaster.
> --------------------------------------------------------------------
> IETF IPng Working Group Mailing List
> IPng Home Page:                      http://playground.sun.com/ipng
> FTP archive:                      ftp://playground.sun.com/pub/ipng
> Direct all administrative requests to [EMAIL PROTECTED]
> --------------------------------------------------------------------
> 
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to