Worthwhile Extensible Internet Registration Data Service
Web Extensible Internet Registration Data Service WG
||Worthwhile Extensible Internet Registration Data Service
Not currently under review
|Send notices to
Internet registries for both number resources and names have
historically maintained a lookup service to permit public access
to some portion of the registry database. Most registries offer
the service via WHOIS (RFC 3912), with additional services being
offered via world wide web pages, bulk downloads, and other
services, such as RPSL (RFC 2622).
WHOIS has never been internationalized. In the absence of formal
specification, ad hoc solutions to signal internationalized
registration data have been adopted and deployed. Providing a
standards-based solution that scales well could minimize further
proliferation of ad hoc solutions.
WHOIS also has no data model: replies are basically just free-form
text. This means that processing of WHOIS output amounts to
"screen scraping", with specialized handlers for every service.
While many of the domain name registries share a basic common output
format, the addition of data elements changes the output
and causes problems for parsers of the data.
The WHOIS protocol does not offer any differential service; it
cannot differentiate among clients to offer different subsets of
information or to allow different access rates to it.
Various attempts to solve the limitations of WHOIS have met with
mixed success. The most recent of these was IRIS (RFC 3981).
IRIS has not been a successful replacement for WHOIS. The primary
technical reason for this appears to be the complexity of IRIS, the fact
that it builds upon many available technologies that in the aggregate
form a complex system. There may also exist non-technical reasons,
but they lie in areas upon which the IETF does not pass judgement.
In recent years, ARIN and RIPE NCC have fielded production RESTful
web services to serve WHOIS data, and each has met with success.
It is widely believed that this simpler re-use of Web technologies
familiar to modern web developers has enabled this success. The
purpose of this working group is to broaden the use of RESTful web
services by achieving simple and common URI patterns and responses
amenable to all number resource and domain name registries.
This Working Group shall determine the general needs of such a
service, and standardize a single data framework. That framework
shall be used to encapsulate objects that could form part of an
answer. The framework shall be for data to be delivered via a
RESTful data service using HTTP (optionally using TLS), and may use
standard features of HTTP to support differential service levels
to different classes of user. The data shall have one mandatory
format, though the working group may consider other optional formats.
The overall effort will be broadly aligned with
a subset of the Cross Registry Internet Service Protocol (CRISP)
Requirements (RFC 3707), but with the explicit additional goals of
producing a simple, easy-to-implement protocol, supporting
internationalized registration data and, specifically for
name registries, capturing the needs of internationalized
domain names in the data model.
As the number registries have more experience with these services
and have found common ground, with their dissimilarities resulting in
more complete working group input documents, the goals of the working
group are to produce standards-track specifications for both number
and name registries using the fashion and pattern of the number registry
input documents, draft-newton-et-al-weirds-rir-query and
draft-newton-et-al-weirds-rir-json-response, as an initial basis.
Work to specify the query for domain name registration data will be
based on draft-sheng-weirds-icann-rws-dnrd.
The Working Group shall determine the general requirements of such a
service, using draft-kucherawy-weirds-requirements as an input document,
and standardize a single data framework. The working group will likely
not seek publication of this draft.
Should the Working Group reach a point where it determines that the
problem of producing a grand unified specification for both numbers and
names appears to be intractable, it will be permitted to divide the
problem into separate tasks and amend its milestones accordingly.