Re: [DNA] DNA and TSLLAO

[Greg Daley <greg.daley@eng.monash.edu.au>]

Hi James,

James Kempf wrote:
> Hi Greg,
> 
> 
>>Just clarifying: if TSLLAO was folded into DNA, then I guess
>>we'd only need the part with RS from an Optimistic address
>>(the part I've already implemented ;-) ,
>>without NS or Directed Multicast?
>>
> 
> 
> I think we could fold the NS usage into DNA too, since testing for
> bidirectional reachability could arguably be part of detecting attachment
> (if you suspect you've lost attachment, this is one way you might want to
> check). Alternatively, one could view the DNA RS/RA as another way to do
> attachment checking, and therefore that NS/NA checking isn't necessary, in
> which case the NS could be dropped.

OK, That sounds reasonable.  The document's structure's pretty logical
due to work by Erik, so it's easy to see where changes would have to be
made in either case.

> With regard to Directed Multicast, I'm not sure I see a general use for
> this, can you give some examples? Typically multicast is used when the IP
> address is undefined.

Yes.  I was wondering if a host may use TSLLAO even though it doesn't
have a full Optimistic DAD implementation.  In that case, unspecified
source addresses would be necessary.

Whether it's worthwhile designing in things for hosts which can
implement one protocol and not another may be debatable though.

Greg