[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: other comments on draft-nielsen-v6ops-3GPP-zeroconf-goals-00. txt

JORDI PALET MARTINEZ wrote:

Hi Pekka,

You mean including in the next version of draft-palet-v6ops-tun-auto-disc
something specific to 3GPP considerations ?

No problem in doing that. Will try to work on this next week, but can't
promise being so fast this time !

Also, we have already some text regarding NAPTR, but we can expand if
required. Any specific suggestion or text that someone want to propose ?


Jordi,

I could send you some text summarizing the NATPR solution,
but this is not what is needed the most at this point in time.

IMHO, we need to understand what are the real requirements
of the different customers (read assisted tunnels, zero conf & 3GPP, potentialy others)
and compare the different solutions with those requirements.

I believe that the 3 customers have the same fundamental need,
i.e. a DHCP based solution won't work (difficult to deploy),
there is a desire to match the underlying topology as well as possible,
and the number of round-trip should be minimize, especially in the 3GPP case.

So, I agree with your conclusion in draft-palet-v6ops-tun-auto-disc-02.txt
that a DNS based solution is the right approach. The point to discuss
is should this be done in the forward or reverse DNS tree? I tend to preffer
doing it in the reverse tree because it matches the physical topology.
The only argument so far I've heard against it is the number of packet exchange necesssary,
(2 instead of 1) however there is a solution to that by padding data in the additional section,
same as what is done for CNAME.

I hope we could find time in D.C. to discuss this further and have the wg come
with a single recommended solution.

   - Alain.