Dyn appreciates the opportunity to comment on the Registration Data Access Protocol (RDAP) Operational Profile for gTLD Registries and Registrars.
Dyn believes enthusiastically in the deprecation of Whois in favor of RDAP. Dyn believes that Whois has long outlived its usefulness and looks forward to the deployment of RDAP as its replacement.
Dyn is aware that there is a Policy Development Process getting started to make a new policy for Registration Data Services (RDS), and that to some extent this proposed Operational Profile is hostage to the current policy. Dyn appreciates that the present policy was developed under circumstances where only the capabilities of Whois were available. Nevertheless, the IETF created RDAP in an effort to correct several deficiencies of Whois. In Dyn’s opinion, the Operational Profile should support those corrections to the maximum extent possible.
Because it is based on HTTP(S), RDAP has the capability to authenticate users. If a server has authenticated the user, then the server can make policy-based choices about what fields to display. Dyn therefore believes it to be advisable that the Operational Profile require by default the ability to authenticate from the beginning. All users, authenticated or not, will of course need to receive the same output, due to the existing RDS policies. But by requiring the ability to authenticate from the beginning, it will not be possible for people to argue during the RDS PDP that differentiated output based on authenticated access will be a new expense: the work of being able to authenticate users will already have been done. Authentication is a key enabling feature of several different policy alternatives, and it should be required from the outset.
RDAP also provides effective referral mechanisms. Given the GNSO’s resolution of 31 October 2013, in favor of the operation of only thick registries, the reason for registrars to need to implement RDAP at all appears to be only that prevailing policy (which is being altered) requires it. This seems wasteful — particularly since the Operational Profile will not provide much of the advantage of RDAP in the form of authenticated access and differential output. It seems it would be better to postpone the registrar obligation until the RDS PDP is completed; at that point, if contracted registrars are still obliged to implement RDAP, they will need to undertake only one implementation effort. Dyn would not especially object to undertaking development knowing that its RDAP obligation would be short-lived, but it objects to undertaking that development without getting many of the important benefits of RDAP.
Dyn is extremely supportive of the requirement to deploy over TLS only, and commends ICANN for it.