Re: [gnso-vi-feb10] The missing part

[Eric Brunner-Williams <ebw@xxxxxxxxxxxxxxxxxxxx>]

Roberto,

The elements of the policy I sent to Mike Zupke earlier today are:

65 Registry Operator (RO) to allow two or more Registry Back-End
Service Providers (RSP) to service provisioning.        

66 RO to provide equal access to RSP(s) from all registrars

67 RO to provide equal access to RSP(s) from all registrants, proxy
and direct

68 RO to provide equal access to RSP(s) for all Registry Services.

69 RO to allow two or more RSP(s) to publish on ports 41 and 80

70 RO to allow two or more RSP(s) to publish on port 53

71 RO to allow transfers of provisioned data between RSP(s) upon
request by the registrar of record.

72 RO to allow transfers of provisioned data between RSP(s) upon
request by the registrant.

What I know about Excel can fit on the head of a pin and leave room
for several angles.

Now you challenge whether this is currently viable. I'm glad to see a
co-chair tasking a proposal advocate to show something. There have
been other opportunities, but this is as good as any other.

For some reason, it could be ignorance, it could be interest, the
economists generating published reports haven't considered the
competition policy issues of opening up the technical provider
(Registry Back-End Service Provider) and contracted party (Registry
Operator) relationship, only the contracted party (Registry Operator)
to contracted party (Registrar) relationship.

Whatever the reason no study of this aspect of competition policy has
taken place, it hasn't.

We know that each of the current gTLD registry operators is
technically capable of running some, or all, of the provisioning,
database management, and publication functions (whois:41 and dns:53),
of the others. We (CORE) are good for 25,000,000 according to our last
audit.

The technical todo list so that EPP payload can indicate a preferred
back-end provider, so that partial zonefiles can be merged, so that
partial views into the database (whois queries) can be merged, are all
tractable. Given that we don't know when the 15th gTLD will be
delegated, it is possible that we could have policy, and the means to
accomplish that policy, for the CNOBI registries before the first new
gTLD application is accepted. A cheery thought.

And, the Registries appear to want the policy for new gTLDs to apply
to existing gTLDs, so unless they don't want competition from each
other, or from all of the registry-capable outfits waiting now, with
the expectation of getting their own monopoly service-provider and
registry operator contract, like the current registries, are likely to
be capable of competing for COM, and NET, ORG, INFO and even BIZ
back-end technical services.

So, who might be interested, other than CORE?

I expect registrars, since if they could steer their EPP add commands
to the lower priced back-end, they could increase their margin until
VGRS moved its price cap down. This could be as much as two or three
dollars per domain year, and in volumes of hundreds of thousands, even
millions, it adds up to real money.

I expect registrants, since they could get lower prices, and could
also self-select back-end operators for some characteristic other than
price, including a desire to end the monopoly regime by the least
disruptive means.

I expect intellectual property holders, since they could replicate
their preference for registrars which meet their generally higher
criteria in selection of the back-end operators who also meet their
generally higher criteria.

Finally, I expect most of the current registries, since all but a few
are captive to a single-provider contract and competition between
back-end providers would, as our economists just told us, lead to
lower prices to them for back-end services.

And who might be opposed?

The Single-Registrant advocates, as they are fixated upon a
top-to-bottom monopoly on the name space(s) they manage to acquire,
and may not think that they can get a better offer from two or more
back-end providers than from one they're stuck with, come riches or
failure, until their string is de-delegated.

Also, any registry with a large volume of registrants and a brand that
means the net and which may be more profitable than it would be if it
had to actually compete out its back-end services to any qualified
provider. I can think of a few names.

There are some issues to iron out. Suppose GoDaddy, with a quarter of
the .com registrations, formed a GoDaddy back-end business unit, and
switched all of its .com registrations to the GoDaddy back-end. We now
have a vertically integrated business unit, from technical back-end to
resellers, with one legal wart in the middle, the 3rd party,
Verisgn-the-Registry-Operator. In no time at all we could have half of
the .com registry on the GoDaddy, eNom, TuCows, and NetSol back-ends.
Is multiple vertical integrations, modulo the persistent legal wart in
the middle, the registry operator, the contracted party, from
bottom-to-top, within the .com market, good, bad, indifferent, or
simply unknown?

Not a lot has changed, really. We can expect the same reaction as was
shown at the Rome meeting, in 2004, where Mouhamet Diop posed the
question of whether competition for backend services would advance
ICANN's competition policy goals, and Ivan Moura Campos ignorantly and
arrogantly treated him like crap, pretending that competition within a
registry amounted to alternate roots [1].

Eric

[1] http://www.icann.org/en/meetings/rome/captioning-board-06mar04.htm