[ppml] Directory Services - Take 2

[Michael.Dillon at btradianz.com]

> > Actually, given the fact that someone whipped together a working
> > prototype of LDAP-based FIRS in a couple of days, I wonder why
> > people continue to chase this overly narrow protocol. Wouldn't
> > it make more sense to develop an XML directory access protocol
> > that provides XML-encapsulated access to directories that are
> > currently stored in LDAP backends?
> 
> LDAP is a protocol.  The back-end for ARIN is what is stored in their 
> relational databases.  Are you proposing that ARIN re-engineer their 
> entire registration system just so they can have an LDAP backend?

No. Change the last line to read:
...currently stored in the backends of LDAP servers.
I'm suggesting that ARIN map it's existing backend database into
an LDAP schema and an XML schema and then serve the data to
end users in the form. And I am suggesting that the IRIS XML
schema should be the model for the encapsulated data that
ARIN puts onto the wire, i.e. they should create an LDAP schema
that is a one-to-one mapping of the IRIS XML schema.

The end result is 3 services using 3 protocols and 2 encodings.

ASN.1 over LDAP.
XML over whois port 43.
XML over REST (which is basically just HTTP with URL-encoded GET queries).

I'm not concerned with the formats or tools used in ARIN's backends
at all. I just would like to see them provide better gateways to
that data, i.e. publish the ARIN whois directory in a manner
that is an incremental improvement over the past and provides the
very real possibility that it may be "good enough" for the next 20
years or so.

LDAP indisputedly has a future. XML indisputedly has a future.
Whois on port 43 is so darned simple that it has a future until
people realise that LDAP or REST are better ways to do the job.
HTTP indisputedly has a future and REST is merely HTTP with the
standard use of URL-encoded GETs and standard documented field
names in an HTML form.

Maybe IRIS has a future. Maybe it doesn't. ARIN can't afford to be
on the bleeding edge. We should be conservative and let the domain
name people sort out the details. If they come up with something
that is as futureproof as LDAP or XML or REST, then we should
consider IRIS at that point in time but NOT BEFORE!

--Michael Dillon