Re: reachability test verification (was Re: Comments on draft-bagnulo-multi6dt-hba-00.txt

marcelo bagnulo braun wrote:
> Hi Jari,
> 
> El 16/11/2004, a las 20:38, Jari Arkko escribió:
> 
>> As for the number of hash verifications, it would seem that
>> you only have to verify those addresses that you actually use
>> for exchanging packets with the peer. That is, if there is no
>> problem you don't necessarily have to verify anything. And
>> if you have a problem, it suffices to test the addresses
>> that you actually end up switching to instead. (I'm still
>> wondering whether the test packets sent to address X need
>> verification of X; there may be some denial-of-service issues
>> if we don't do that.)
>>
> 
> The question would be: do we do the reachanility test first or the HBA 
> verification first? is this what you have in mind?

That was the question I had. It seems that it would be best
if you could verify HBA/CGA first and only then send a packet.

--Jari