From: Ross Wm. Rader Sent: Friday, 5 October 2001 2:10 AM To: jo.lim@auda.org.au Subject: Tucows Comments on: "dotAU Domain Administration Ltd: Registry Technical Specification" I write to you on behalf of Tucows Inc., a registrar in a number of top-level domains (generic and country-code). In addition to our registrar services, Tucows also develops and sells a software platform called OpenXRS. OpenXRS is a registry services management platform that is used by a number of TLDs today. In creating OpenXRS, Tucows carefully considered the same issue that auDA faces today - using an existing protocol, creating an interim protocol or capturing and implementing the essence of an emerging protocol. I would like to share some of our thinking behind our ultimate choice. Initially, our thinking was largely limited to the first two choices. The Network Solutions RRP protocol specification was reasonably mature and had been successfully implemented by dozens of Registrars. Unfortunately, we deemed that it was neither extensible nor scalable enough to suit our long term business objectives. It appears that this thinking has been largely reinforced by the industry. With this in mind, we undertook the development of an interim protocol that would satisfy our objectives. This interim protocol, dubbed "xRRP" (extensible RRP), mimicked many of the behaviours and actions of the RRP but was based on XML. There were also several other key additions to the protocol that provided, what we felt, was a much superior solution. However, during this period, the IETF Provisioning and Registration working group was struck and the EPP emerged as a highly capable competitor to the existing RRP and the Tucows xRRP protocols. Realizing that EPP's potential was still largely undefined, being an emerging standard, we felt it was most appropriate to abandon our proprietary effort and embrace the EPP initiative. At this time, doing so was a risky proposition. EPP had no track record, the specification was subject to change at the whim of the working group and there were few, if any, experts to draw upon. Given this, it was felt that our best option was to settle on a definition of EPP that we were comfortable with, implement that and then once the standard was ratified, immediately move to adopt the final standard. This strategy has served us well. In adopting the pre-standard specification, we have not only afforded ourselves the opportunity to build expertise internally regarding the nuances of the protocol, but we have also been able to accelerate adoption. Adoption is important on two levels - users and developers. From a user standpoint, over 80 registrars have successfully implemented the EPP protocol in their back-end registrars systems. Developers have also embraced the challenge by participating in the Tucows sponsored EPP Registrar Toolkit initiative. The goal of this effort is to provide registrars with a universal tool-kit that will work with all EPP based registrars. Further, the software developed through this initiative is open-source and available for modification and use by all. We would urge auDA to consider their options carefully in this regard. As many of the other comments received by auDA indicate, implementing a new interim protocol when there are several other, more defined specifications available, is a risky and expensive proposition. Adopting an EPP-based solution, however, would provide auDA with a stable and supported base from which to launch your registry efforts. It will also provide a clear and well-defined migration path as the working group further refines the specifications. Further, auDA will also benefit from the existing operational expertise of the registries and registrars already using EPP-based systems. This knowledge would have to be created if auDA chooses to continue down the iRRP path. We hope that this information is useful in your deliberations and wish auDA, and Australia generally, the best of luck in this exciting endeavour. Sincerely, Ross Wm. Rader Director, Innovation & Research Tucows Inc.