Quoting Ron Stark on Monday May 17, 2004:
| Kim, you've omitted the issue of return on investment that a registry might
| have made.  If the investment was $1000 to set up and the running costs
| $1000 per year, then a 20% "markup" is high.  If the entry cost was $5M, and
| the running cost $2M per year, then a 500% is too low.
| 
| Much of the value in a business lies in the resale value of the business
| itself.  In the registry's case, if they lose the next tender their entire
| investment to set up is lost - the business itself probably won't be sold to
| the next successful bidder.  On the other hand if they ARE successful,
| domain name prices should arguably drop dramatically because the registry no
| longer has an initial investment to recover, just running costs to meet.

Whilst there is clearly some cost in establishing a registry system,
it need not be too high. In the tender process, auDA sets some minimal
levels that a registry must comply with, but other than that there is
not much to it.

The technical foundation of a registry is not that complex. It is a
glorified database, with some unique, but relatively straightforward,
interfaces to it.

Alot has changed in the time since auDA let out the contract to run the
current registry operation. The protocols have matured, and there are
free software platforms, as well as healthy competition from proprietary
vendors, that can run EPP-based registries. In much the same way that
free software platforms run the DNS servers themselves, the registry
systems are becoming commoditised.

It is also worth noting that the investment in building a registry can
be transfered to other registries - it is not all lost if you lose a
rebid. AusRegistry, as far as I know, already runs registries in other
countries using the system they built first for .au. So, there are still
market opportunities if you don't use it for the same client, it is
reusable and not a complete writeoff.

kim