Re: URN Resolution Paths Considered Harmful

Mark Madsen (
Thu, 6 Jul 95 16:07:19 BST

From: Mark Madsen <>
Message-Id: <>
Subject: Re: URN Resolution Paths Considered Harmful
Date: Thu, 6 Jul 95 16:07:19 BST
In-Reply-To: <>; from "Larry Masinter" at Jul 5, 95 5:21 pm

Larry wrote:

> Should someone be able to replicate a URN->URL service? I.e., is
> there a way to ask the URN->URL service: "What URNs do you resolve
> and what are their URLs?"

Let me answer this with a question.  Can you think of an example of a
service that should _in_principle_, *not* be replicable?  And there
are many reasons why you would want to be able to replicate (and
migrate) services automatically.

That's not to say that it's up to a standard to say how replicable a
service needs to be, or how it is to be replicated.  Intuitively, you
want replication to be invoked through some level of management

The example question you pose is not quite the same as the replication
question, though.  Consider a URN resolution service with the
following interfaces to an unnamed URN resolver object (in something
like CORBA-IDL):

Interface URN_Resolver_Resolution{
    array-of-string Resolve(in string urn_to_be_resolved);

Interface URN_Resolver_Enquiry{
    resolutionTable EnquireAll(in string name_of_turkey_who_requested_this);

Interface URN_Resolver_Management{
    void Replicate(in string dest_IP_address, in string name_of_clone);
    void Migrate(in string dest_IP_address, in string name_of_new_me);
    void NukeTurkey(in string turkey_name);

This should make the distinction clear.

> Are there any privacy issues here?

Yes.  But if you only allow access to replication methods through
management interfaces, you can contain those issues as tightly as they
are contained in the physical world.

Mark Madsen: <> <URL:>
Information Services Framework, The ANSA Project, APM Ltd., Castle Park,
Cambridge CB3 0RD, U.K.  <URL:>;  <>
Voice: +44-1223-568934; Reception: +44-1223-515010; Fax: +44-1223-359779