[ga] Date: Wed, 9 Jan 2002 15:53:53 -0500

["Elliot Noss" <enoss@tucows.com>]

I wanted to take a bit of a "clean-sheet" approach to this discussion as the
points I wish to communicate cut across a number of different threads on a
number of different lists.

There are really four topics I wish to raise as follows:

- Who has the "right" to deal with expired/expiring names?
- The mixing in this discussion of two seperate issues, registry load and
the allocation of expired/expiring names;
- The inefficiency that results from any flat-price solution;
- A way forward.

Who has the "right" to expired/expiring names?
-----------------------------------------------
This is an issue that creates an interesting sub-text to this entire
discussion, yet has not been fully examined. There are three potential
claimants for this right and three possible states for these names.
Claimants include registrants, registry and registrars. States include
unexpired, expired in the grace period and expired o/s the grace period.
There are two things that are clear. First, that no one party is clearly
entitled to stake a claim. Registrars are limited by the terms of 3.7.5 of
the RAA which require that names be put back in the pool if not renewed. The
registry is limited by its role as monopoly technical supplier and by the
Registry agreement which entitles it to a fee for the services it is
contracted to perform and confers upon it no property rights beyond that.
Registrants are limited by a number of practical issues including their
limited rights in a name and their diffuse nature.

What is clear to me is that this IS NOT a function of a registrars terms of
service, nor is it an inherent right contained in the registry agreement.

At the same time we must keep our eye on the fact that the role of
registrars and registry is, most purely, to efficiently administer the
allocation and provisioning of domain names. This means that the best
approach is the one that puts names into the hands of those who would put
them to the most use. Names in the hands of those who most desire them will
lead to a fuller utilization of the Internet, more value for users and more
revenue for registrars and registries.

This point should not be seen as at odds with an egalitarian (as opposed to
equitable) view of domain names and first-come-first-served ("FCFS"), but I
realize this is a point that would be the subject of much debate. What is
interesting today is that we are at a unique time and place in the history
of the DNS which makes this less contentious. We have one extremely mature
namespace in .com/.net/.org. It is almost certain to be the largest
namespace throughout the lifecycle of the current DNS. It also has a
secondary market that is more evolved than any other will ever be. We also
have the recent introduction of new gTLDs that provide a fresh supply of
names. This means any solution effected can be tailored to the current
circumstance.

We must also remember that there are two groups of registrants that we must
consider, current registrants and potential registrants. They have distinct
interests. Current registrants have rights around their existing names, both
in terms of security from losing a name through inadvertance and in excess
economic value. Potential registrants benefit from being able to efficiently
obtain names that are currently owned. With the introduction of new gTLDs
potential registrants have, and will continue to have more and better
alternatives. The maximum value in the secondary market exists right now,
today.

At the end of the day the competing claims of registries and registrars are
likely subordinate to those of registrants. Accordingly, any solution should
start with this underpinning.

Registry load and the allocation of expired/expiring names
-----------------------------------------------------------
It has been noted by a number of people in this debate, and has been my
position for many months, that the issues of registry load and the the
allocation of expired/expiring names are being mixed together unnecessarily.

I wish to add my voice to the chorus saying that these issues are related
only remotely, almost accidentally. There have been a number of very good
suggestions as to simple steps the registry could take to lessen the load.
There are a couple additional points worth noting here. First, the current
solution is no longer broken. While I am not a fan of the status quo, the
registry has weathered the storm and there seems to currently be no
appreciable impact on our day-to-day business (which was not the case a
short time ago). An additional measure or two (a modified check command,
additional, transparent compliance, all names dropping in real time and a
published drop list are the easiest and most effective IMHO) would make this
a non-issue.

It is worth noting my personal dealings with the registry on the question of
load have been positive and I was impressed with their genuine desire to
solve the issues at hand.

This last point leads me to feel comfortable that these issues are not being
presented as being directly connected and can be dealt with seperately. I
would, of course, love to hear Chuck confirm this.


The Inefficiency of flat pricing
---------------------------------
The current market for domain names is characterized by flat-priced supply
and variable-priced demand. I do not take a politial position on this, I
merely note it as observation. This market inefficiency (again observation
not position) has lead to the existance of a robust secondary market. It has
also lead to a significant amount of the current CNO namespace sitting
unused.

I have strong reservations about any solution geared at the expiring market
that magnifies that inefficiency. By definition it leaves money on the table
and leaves demand unfulfilled at the same time. The worst of both worlds.
Ideally we could find a solution that was able to create a robust, efficient
secondary market which would benefit registrars and the registry, but if
done properly would most benefit registrants.

A suggested way forward
------------------------
Unfortunately, for me, the existing WLS proposal is not acceptable. The
inefficiencies are large and the economics are miles away from either fair
or realistic. With significantly re-worked economics it could be an
acceptable interim step, but I am not sure we need an interim step,
especially given the decoupling of the registry load issue.

It seems to me there is a way forward that addresses all of the above
issues. I would suggest two important modifications to the existing Peter
Girard proposal. An unlimited bidding period and the bulk of the fees going
to existing registrants rather than registrars.

All names should be available to "bid" on at any time. A "bid" by a
prospective registrant would require an administrative fee collected by a
registrar, shared with registry and would be available for acceptance by the
existing registrant at any time. A sucessful transaction would lead to a fee
to both registry and registrars. An example:

- Potential registrant places a bid of $150 on abcd.com and for doing so
pays a non-refundable administrative fee to registrar x of $10 and in turn
registrar pays registry $5;
- Original registrant is made aware of his ability to "transfer" the name
and any unexpired term to a potential registrant for $120;
- If original registrant decides to accept he contacts the existing
registrar of record and informs him of his desire;
- If the registrars are different the $30 transaction fee is split 1/3 each,
if the same than the split is equal between registrar and registry;
- The fee would be a % of bid, capped at a relatively low number ($30?).

To be clear, this is described in very brief terms and would need
significant rounding out as well as a champion so please work the principals
not the specifics.

This solution would provide significant benefits to everyone involved in
both an equitable and palatable fashion. It would also keep both registrars
and registries in their role of market makers not market participants and
would create a level of efficiency that would lead to increased revenues,
increased registrant satisfaction and, perhaps most importantly, maximized
use of the namespace.

The original administrative fee would/should act so as to deter nearly all
wasteful behaviour. Technically, it need be no more complicated than an
interface between the SRS and EBay's open APIs (full credit here to Joyce
Lin in Montevideo, if only we would have listened to you then!). It could be
completely done by the registry, by the registry with technical partners,
Peter could do it, or it could be put out to a completely new tender
process. I know we would love to build it (but to be clear we are not
throwing our hat in the ring whatsoever). I know perhaps fifty people
reading this message would love to build it. That kind of excites me,
thinking of all of the extremely capable people who could nail this
technically. Innovation would abound if we let it.

The single largest beneficiary would be the registry. Ok by me. The largest
cumulative benefit would accrue to registrants. Again, that works for me. I
think this would be an absolutely elegant outcome for everyone.

Regards

Elliot Noss
Tucows inc.
416-538-5494

--
This message was passed to you via the ga@dnso.org list.
Send mail to majordomo@dnso.org to unsubscribe
("unsubscribe ga" in the body of the message).
Archives at http://www.dnso.org/archives.html