Re: new port for DNS

Alexander Dupuy (dupuy@smarts.com)
Wed, 21 Jun 1995 22:25:03 -0400


Date: Wed, 21 Jun 1995 22:25:03 -0400
From: dupuy@smarts.com (Alexander Dupuy)
Message-Id: <9506220225.AA28953@just.smarts.com>
To: ietf-lists@proper.com
Subject: Re: new port for DNS
Cc: uri@bunyip.com

> Just for clarity, are you proposing a domain name system ending with
> ".urn", such as "proper.urn"? If so, who would assign the second-level
> names? The InterNIC? And, if so, what about people outside the US who want
> names?

Yes, there would be a .URN domain.  I would guess that IANA, rather than the
InterNIC, would be responsible for assigning the second-level names.  The
second level domains would be the names of the URN naming authorities, so that
for example ISBN-based URNs could be resolved under .ISBN.URN (this might need
special nameservers that algorithmically generate CNAMEs, like the
IDDD.TPC.INT nameservers do).  For new-style URNs, each of the competing
proposals could get a second-level: e.g. OID.URN, and whatever else.  The
naming authorities would then delegate pieces of their subdomains according to
whatever policies and procedures they had set up.

> Another similar option is to allow anyone to be a URN service. I earlier
> mentioned the fictional URNsRUs, whose US-based domain name might be
> urnsrus.com. The naming authority in a URN that was through this company
> might look like:
> proper.urnsrus.com
> 3.5.9.2.100.2.urnsrus.com  (Look, Michael: OIDs!)
> and so on.

This is actually a pretty good example of why using existing DNS names is a
bad idea.  Presumably URNsRUs will use urnsrus.com for their own hosts.  This
will cause security concerns if they wanted to delegate proper.urnsrus.com
URNs to nameservers at proper.com - making URNs containing urnsrus.com much
less flexible than URNs using just OIDs.  Not to mention the problems that
might be caused if a company like O'Reilly started a URN resolution service
using ora.com, but then sold off the service to America Online or OCLC.

The more I think about it, the more convinced I am that if URNs are to be as
persistent as possible, that they should be numeric (or alphanumeric codes
like the LoC numbers or British/Canadian postal code system).  If you use
human-readable names like "proper" or "ibm" people will get emotional and/or
possessive about them, making it much harder to prevent the URNs containing
them from changing over time.

> As you can tell, I'm still against a central naming hierarchy, even if it
> is ".urn". I don't trust the namer delegators to give them away freely and
> fairly, I don't trust them to resolve them freely and fairly, and so on.
> The current DNS still sounds pretty reliable to me.

I'm against a single central naming hierarchy; but I think having a few central naming hierarchies should be okay.  If you aren't willing to trust IANA to
require reasonable subdelegation arrangements from naming authorities, why are you willing to trust the IETF standardization process to come up with a URN
protocol?

@alex
--
inet: dupuy@smarts.com
Member of the League for Programming Freedom -- write to lpf@uunet.uu.net
GCS d?@ H s++: !g p? !au a w v US+++$ C++$ P+ 3 L E++ N+(!N) K- W M V- po- Y+
     t+ !5 j R G? tv-- b++ !D B- e>* u+(**)@ h--- f+ r++ n+ y+*