Re: [DNA] MLD JOIN Questions?

[Christian Vogt <chvogt@tm.uka.de>]

Hi Sathya, hi Greg.

> The host SHOULD send the MLD group join report for the solicited
> node's multicast group when it becomes optimistic.  [...]

Note that Nick mentioned that he'll modify the OptiDAD draft so that the 
optimistic address will be available for use even before the initial NS, 
and thus the MLD report, is sent.  (See thread starting at [1].)

This would allow an RS with a TSLLAO option to be sent even before the 
mobile node joins the solicited-node multicast group corresponding to 
the optimistic address.

Does this have an impact on the DNA Protocol 3?

Regards,
- Christian


[1] http://www1.ietf.org/mail-archive/web/ipv6/current/msg05777.html

-- 
Christian Vogt, Institute of Telematics, University of Karlsruhe
www.tm.uka.de/~chvogt/pubkey/


Greg Daley wrote:
> Hi Sathya.
> 
> I've written some answers to your questions, in the rough form of a
> text proposal:
> 
> 
> Sathya Narayanan wrote:
> 
>> Before I forget, I just wanted to capture the questions and post
>> them. The questions are:
>> 
>> 1) When should the host do MLD join solicited node multicast
>> group(s)? - After marking the IPv6 addresses as optimistic, or -
>> Just before sending RS message or - someother time
> 
> 
> The host SHOULD send the MLD group join report for the solicited
> node's multicast group when it becomes optimistic.  Note that this
> may precede transmission of the DAD NS probe by some time, although
> the optimistic node's ability to receive DAD NS's is limited until it
> does so.
> 
> In any case, the MLD report message MUST be sent before transmission
> of either an RS, otherwise responding routers may not be able to
> resolve the host's link-layer address.
> 
>> 2) Do we have to make any recommendation regarding MLD joins when
>> layer-2
> 
>> hints are received whether there is a link change or not? (due to
>> the
> 
>> presence snooping switches)
> 
> 
> A host SHOULD always join MLD groups, after it receives a link-UP
> hint on a new point-of-attachment.
> 
> Until it does so there is no guarantee that it can receive packets 
> addressed to its solicited nodes' addresses.  This means that it may
> not be able to see DAD NS probes or neighbour resolution attempts 
> until it has done so.
> 
> 
> Comments?
> 
> Greg