Re: [homenet] tunnels as way to disambiguate .local

[Brian E Carpenter <[e-mail]>]

On 08/08/2012 15:09, Andrew Sullivan wrote:
> On Wed, Aug 08, 2012 at 01:03:11PM +0100, Brian E Carpenter wrote:
>> It certainly is. But see http://tools.ietf.org/html/draft-carpenter-referral-ps
>> especially section 4.2 "FQDNs are not sufficient".
> 
> I thought one of the things we were trying to do was to address
> exactly the failure modes in that section of the I-D?

Yes, I guess that's right. I hadn't looked at it that way.

> Perhaps I'm being naive, but I've been working from the assumption
> that, if you want to talk to something on the Internet, you need an
> unambiguous way to identify it.  Historically, the best we've had for
> that has been the DNS, because it provides a layer of indirection so
> that you can have stable identifiers in the face of changing IP
> addresses.  

What's actually happened (IMHO) is that people trying to solve the
reference problem have come up with additional ID spaces to solve their
particular problem - often hanging them off a trouble-free FQDN to
get the top level of uniqueness for free. But then they have to invent
things like STUN, ICE or supernodes to get connectivity to those IDs.

> Given the way the relevant markets have gone, it turns out that DNS
> names are rather harder to administer and use for ordinary end users
> than we might like.  But there's no reason that has to persist, and it
> seems to me that if we're going to solve the problems people have
> using homenet-type resources on the global Internet, then solving the
> DNS piece in a user-friendly way is going to yield greater benefit
> than alternatives like ginning up some trick to make mDNS names
> sometimes work outside their natural context.

I'm inclined to agree. If we phrase it in terms of specifically
eliminating the problems in my draft, I'll even be happy about it ;-).

   Brian

> 
> Best,
> 
> A
>