Re: [Mip6] draft-dupont-mip6-dhaadharmful-00.txt

[Alexandru Petrescu <[e-mail]>]

Francis, please allow me slight remarks, don't take as corrections.

Francis Dupont wrote:

>  In your previous mail you wrote:
> 
>    > => we agree: current DHAAD design is not stupid, it has only two real
>    > problems:
>    >  - the DHAAD idea itself is in danger because HA assignment is better
>    
>    By "HA assignment" you mean HA Switch message and draft?
> 
> => no, I mean a HA assignment mechanism in general, including
> (i.e., mainly :-) any still to be invented mechanism.
> HA switch message is fine but is clearly *re*-assignment.
> 
>    If yes, that
>    draft alone does not help for bootstrapping IMHO.  If DHAAD can be made
>    to work with non-anycast addresses then DHAAD+MPD is a good idea to do
>    bootstrapping simple (MPD is Mobile Prefix Discovery).
>    
> => I don't understand why MPD is for when you have HA discovery/assignment
> (the mobile prefix is simply the prefix of the HA address)... I should
> have missed something, for instance there is more than one MP?

"MP" could be misinterpreted as "Mobile Prefix" which isn't.

MPD is more like: Mobile       Prefix Discovery, a prefix discovery
performed by a mobile.

There's only one MPD.

DHAAD doesn't provide a prefix length field in DHAAD Reply, so the MN
can't know it.  MPD is the equivalent of local RS/RA for a MN that is
away from home.

For bootstrapping, MN could do DHAAD to get the HA's full address and
then interrogate that HA with MPD Sol and get Response, and then call
the stateless address autoconfiguration functions easily.

HA-Switch and DHAAD+MPD could also be used when the home link is
re-numbered.

>    >  - the current mechanism has some operational issues, mainly its lack
>    >    of reasonably easy security and the spurious but common ICMP filtering.
>    > My draft is about the second but the first is more critical and drastic...
>    
>    I don't see anything critical or drastic in HA assignment being proposed
>    as better than DHAAD.
> 
> => the question is drastic, i.e., if the answer is yes all discussions about
> the best DHAAD mechanism are lost bits...

I agree.

[... snip because I don't know]

Alex


_______________________________________________
Mip6 mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/mip6