Re: A philosophical question about large directory structures

At 08:19 23/06/99 +0200, Harald Tveit Alvestrand wrote:
>the example is not terribly encouraging; today *everyone* is grabbing
>their own IP addresses, but they do it from Net-10 and expect NAT to solve
>all their problems.
>
>The equivalent directory gateway *could* exist.

I think comparison with the Internet and IP is interesing, and might even
be instructive.  Some random thoughts:

- the Internet model of linking together free-standing networks that use
common protocols would seem to be applicable to directories.  The resulting
directory would, I think, be not so much a single unified global directory
but a less organized network of directories that can work together to
provide something approaching a global information infrastructure.

- I think that, like IP networks, directories will come about if they
justify their existence when used in isolation;  interconnecting them would
come later, if the technical design and implementation allows that.

- at least directory naming doesn't have an address space restriction.

- I don't think directory services have the same architectural requirement
for end-to-end transparency that exists in the Internet, so namespace
gateways, if needed, might present less of a problem.  But it would still
seem preferable to avoid them altogether, for reasons of relative simplicity.

- I think a big challenge area will be directory schema:  I think these may
play an analogous role in a directory infrastructure that protocols play in
the network infrasructure.  I think schema gateways, if needed, may prove
more problematic than namespace gateways (c.f. protogol gateways and
address translation).

>I think no global infrastructure (like namespace) can usefully exist
without a
>global coordination and a global "something" that works.
>I sometimes hope I'm wrong.....

In a different context, I recently gave a little thought to this idea of
global name space and central cordination.

Broadly, I tend to agree with your conclusion w.r.t. directories.  But
consider:  many problems become considerably less difficult if one accepts
that a correct answer is delivered with some probability less than 1, but
which can be made arbitrarily close to 1.  This idea is used explicitly in
a recently published draft for identifying content media feature
collections:  an identifier with some randomized (but repeatable) property
that is used to serve as an identifier.  If the identifier is long enough,
and sufficiently independent of other germane data ("random"), then its
uniqueness can be asserted with a probability arbitrarily close to 1.
(This assumes that uniqueness is the reason for central management of the
namespace -- there may be other considerations.)  (The draft mentioned is
<draft-ietf-conneg-feature-hash-02.xt>.)

#g


------------
Graham Klyne
([email protected])