On Wed, 28 Jul 1999 [EMAIL PROTECTED] wrote:
> > Insofar as you are commenting upon this, you seem to have missed my
> > narrow technical point. Roberto Gaetano asserted that an IP address
> > uniquely identifies a domain name. This is not true.
>
> Speaking about missing the points, I come back to this issue because you
> raised it again. The point I was making in my original note was completely
> incorrelated with whether an IP address uniquely identifies a domain name or
> not.
>
> You picked the sentence:
>
> > I find the analogy with the phone system (as you present it) not fully
> > applicable, as the phone number is a "key" in the system, and
> > therefore
> > unique due to the way that the system is built, while the
> > domain name is an
> > "attribute" of the unique key (the IP address), and therefore could be
> > duplicated.
>
> and made a whole story about the "unique key", interpreting this as if I
> said that the relationship domain name - IP address is unique (which I did
> not say).
Sorry, wrong again. I interpreted as though you said the IP address is
the unique key, which you did. In relational database jargon, this means
that given an IP address, you know what the domain name is. The fact is
that you don't.
> The example was aimed at showing that while the telephone numbers are
> unique, domain names are not if we allow multiple independent roots.
Well, if we are kind and accept that that is what you intended to say,
you didn't say it.
However, it still is not necessarily true. Some time back in this
thread I pointed out that Internet routing has in effect thousands of
independent roots, the autonomous systems (ASs) which in combination
are the Internet. Each AS has its own separate routing policy, each
decides independently which routers to announce and how to announce
them.
You imply that a DNS with multiple roots must necessarily have multiple
overlapping name spaces. This is not true.
> Therefore, while if you pick up the phone and call a number you will get
> *always* to the same person,
What?! Not in the universe that I live in. If you call 192 (directory
inquiries/"information") in the UK, you are very unlikely to get the same
person. I don't believe that you even get the same city.
Of course if you call 999 (emergency services) in three different towns
in Britain, you will get a different person working for a different
organization each time.
That is, neither the telephone system nor the domain name system works
as you claim it does.
> in the proposed system you type a domain name
> and you will get to a different domain *depending on how your system is
> configured*.
Sometimes yes, sometimes no.
> And this result is independent from uniqueness or not of the relationship
> domain name - IP address.
Untrue. If the domain name - IP address mapping were 1:1, you would be
guaranteed to get the same IP address (which is what I suppose you mean
by the term "domain" in this sentence). The relationship matters.
> I am honored that you broke your long silence to reply to a message of yours
> truly, but I would have appreciated an opinion about this problem of
> substance, i.e. on how to guarantee certitude to reach the business you want
> to address on the Net via its domain name under a multiple root system
> (independently from the configuration of your computer and/or the choices of
> your ISP).
You begin with a simple mapping from IP address to domain name. You
have drifted towards a (domain name --> business) mapping. I think
that really this is further than you want to go. It's just too vague.
What you want to say is something like this: the question is how to
guarantee certitude that where the holder of the domain name x.y.z
requires that it resolve to an IP address A.B.C.D (or set of such
addresses), this IP addresss (or set thereof) be returned by all
resolvers.
Does such certitude require that all of the Internet's autonomous
systems swear fealty to ICANN? Absolutely not. Does it require
that ICANN have the power to impose taxes on the Internet? Absolutely
not. Does it require that there be a single primary root name server,
a.root-servers.net, that all of the other servers in the world defer
to? No. Absolutely not.
There are many examples of communities that cooperate voluntarily
to agree on unique identifiers without surrendering to some
central authority. Banks, for example, have unique bank numbers
that everyone recognizes. This does not mean that all banks in the
world are subservient to some central authority. On the contrary:
the banks fund some obscure committee somewhere that assigns the
numbers. If Joe's Bank, Garage, and Beauty Shoppe chooses to
ignore this committee and use the same bank number as someone else,
all that will happen is that clearing houses of various types will
have to process their paperwork (cheques, wires, etc) by hand. Or
they will refuse to do so.
Nothing more. Nothing less.
We don't need anything more than this level of coordination in
the Internet. That is, we need IANA. We need someone to formally
say that the next protocol number after 17 is, yessir, 18.
We don't need ICANN as the Great Regulator. We don't need ICANN
as a Global Taxation Authority. We don't need a vast grey bureaucracy
in the Southern California desert telling the whole world what to do.
We don't need ICANN's unelected, unknowledgeable, uninformed, and
generally rather useless board demanding money from us with menaces.
What we need is the sort of thing that IANA used to be: a small
group of intelligent, earnest people with good intentions who help
us all to voluntarily cooperate for our mutual benefit.
> > Sometimes a name
> > corresponds to many IP addresses (as in round-robin DNS) and sometimes
> > an IP address corresponds to many names (some Web servers permit many
> > names to be associated with one IP address). The domain name system
> > is not really like the telephone system.
> >
>
> So what?
>
> Regards
> Roberto
>
> P.S.: Of course it is not. If they were identical we wouldn't need both ;>)
"Similar to" is not the same as "identical to". What I said was that the
domain name system is not really similar to the telephone system. In fact
they are very very different.
--
Jim Dixon Managing Director
VBCnet GB Ltd http://www.vbc.net tel +44 117 929 1316
---------------------------------------------------------------------------
Member of Council Telecommunications Director
Internet Services Providers Association EuroISPA EEIG
http://www.ispa.org.uk http://www.euroispa.org
tel +44 171 976 0679 tel +32 2 503 22 65
