Re: Welcome!

[Alex Bligh <alex@alex.org.uk>]

--On 17 May 2004 16:47 +0000 Paul Vixie <paul@vix.com> wrote:

> for the record, isc would happily create "e164.isc.org" and add
> delegations at the 100- or 1000-block to anyone who faxed us a phone
> bill, if there was

Hmmm... the trouble with this is "you and how many others". Whilst E.164
numbers are (in essence) unique, i.e. there is a canonical albeit
distributed "database", what you are trying (for good reason) to do is
generate another database of "who has what E.164 number" (for a subset of
all E.164 numbers), because (in essence) you see the ITU's/USG's
role/action (or rather lack thereof) as problematic. Who is to say someone
else won't have the same opinion? I can't help but be drawn to a comparison
with the "Alternate Root" crowd.

Perhaps I look at this whole thing backwards and think that I should
be phoning Paul Vixie, and that's paul@vix.com in the first place.
Doing a lookup to E.164 number and then inverting back to an IP address
is wierdness I shouldn't need to do other than for legacy reasons. Thus
the E.164 numbers I'm likely to call are precisely those which are
least likely to be in Paul's registry (which seems to be end-user
initiated) as those who are in Paul's registry I can likely call using a
domain name and appropriate SRV/NAPTR record anyway.

So it seems to me the important thing is either for the underlying
operator to give an ENUM response (or a default ENUM response) so I
can terminate calls to non-VoIP people, or for there to be some sort
of competition (multiple routes) to the terminating line.

And thus, as such, I don't think you are actually solving the right
problem. If I want to terminate a call on +1.650.423.1301, I don't really
care what telco that's on. Or indeed who has the phone bill for the number.
Just like if I want to reach 192.228.79.201, I don't really care at the
routing level who owns the address block. All I want to know is who will
accept traffic for +1.650.423.1301 and (possibly) under what terms they
will bill me for it, what QoS etc. I do that on a good day with BGP at the
network level. It doesn't tell me the RIR entry and who owns the block, it
tells me how I get there. [yeah sure we goofed when we designed BGP in
that there is in essence no security, or, being polite, only transitive
security between BGP speakers and no end to end security, but that's
nothing to telcos who in effect have designed interdomain routing with
static routes]

How is all this relevant?

Well, if I use Paul's model, then for the first two years the only entry
starting +1.650 I'm ever going to see is for a Mr P Vixie Esq. on a rather
more specific address (ok, well may be a few more). If I could, on the
other hand, do some form of lookup for +1.650.xxx.xxxx and found Mr P Vixie
would terminate calls to +1.650.423.13xx for free, and a host of telcos
offered +1.650.xxx.xxxx at various different rates, possibly including
zero, I'd be sending a lot more calls that way.

I'm not sure this is something DNS can effectively do (probably, it's
been (ab)used for enough else). But if it is, I'm sure Paul's going
to be the one to tell us how.

Alex

--
To unsubscribe send a message to voip-peering-request@psg.com with
the word 'unsubscribe' in a single line as the message text body.
An archive is at <http://psg.com/lists/voip-peering/>.