draft-skwan-utf8-dns-06.txt

bind-3.2.

INTERNET-DRAFT                                             Stuart Kwan
                                                          James Gilroy
                                                          Levon Esibov
                                                       Microsoft Corp.
                                                              May 2001
<draft-skwan-utf8-dns-06.txt>                    Expires November 2001


     Using the UTF-8 Character Set in the Domain Name System

Status of this Memo

This document is an Internet-Draft and is in full conformance
with all provisions of Section 10 of RFC2026.

Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups.  Note that
other groups may also distribute working documents as
Internet-Drafts.

Internet-Drafts are draft documents valid for a maximum of six
months and may be updated, replaced, or obsoleted by other
documents at any time.  It is inappropriate to use Internet-
Drafts as reference material or to cite them other than as
"work in progress."

The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt

The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.


Abstract

The Domain Names standard specifies that hostnames are represented 
using the ASCII character encoding.  This document expands that 
specification to allow the use of the UTF-8 character encoding, a
superset of ASCII and a translation of the UCS-2 character encoding.


1. Introduction

The Domain Names standard [RFC1123] specifies that hostnames are
represented using the ASCII character encoding.  This document expands
that specification to allow the use of the UTF-8 character encoding
[RFC2044], a superset of ASCII and a translation of the UCS-2
character encoding.

Interpreting names as ASCII-only limits the utility of DNS in an
international setting.  The UTF-8 character set includes characters
from most of the world's written languages, allowing a far greater
range of possible names and allowing names to use characters that are
relevant to a particular locality.  UTF-8 is the recommended character
set for protocols that are evolving beyond ASCII [RFC2130].

Expires November 2001                                         [Page 1]

INTERNET-DRAFT                  UTF-8 DNS                     May 2001


This document defines the technology for a richer character set in
DNS.  This document specifically does not define policy for the
characters allowed in a name when used in a particular application.
For example, some protocols place restrictions on the characters
allowed in a name


2. Protocol Description

2.1 Components and roles

Before the description of the protocol itself authors feel a need to
clarify which components are involved in processing the hostnames and
describe the usage of the hostnames by these components. The following
list contains such information.

User.
User could be a human or application. Its role is to specify (also
known as "write") and retrieve (also known as "read") the hostname to
and from an application. The examples of such operations include
typing the hostname, writing it on a touch sensitive screen, reading
the name from the monitor, listening to a voicemail, etc...

Application.
Application's role is to
- process the hostname specified by user or other local or remote
  application.
- return to the user (for example display on a monitor screen) the
  hostname returned by DNS resolver.
- call DNS name resolution APIs to request resolver to perform the
  name resolution

Resolver.
Resolver's role is to
- process the name resolution requests from an application and submit
  appropriate DNS query to the DNS servers
- process the response from a DNS server and pass the response to the
  Application.

DNS server.
The role of the DNS server is to store and maintain the DNS data,
process the updates to its database, update the replica copies of the
databases and perform the DNS name resolution through responding to
the DNS queries.


2.2 Protocol details

This section describes the modifications (if any) to each of these
components and interfaces between the communicating components.



Expires November 2001                                         [Page 2]

INTERNET-DRAFT                  UTF-8 DNS                     May 2001


2.2.1 Users

No modifications to the users are proposed in this document. At the
same time support of this protocol by other components specified later
in this section may enable users to start using in hostnames
characters from wider set than one specified in [RFC1123].


2.2.2 Interface between users and applications

User may use any character set or multiple character sets supported by
the particular application. Specification of the allowed character
sets supported by an application is outside of the scope of this
document. The decision on which characters sets can be used to allow
user to input and retrieve the hostnames is left to the implementers
of the particular applications unless a protocol underlying specific
application specifies the supported characters set. Thus this protocol
does not affect the interface between users and applications.


2.2.3 Applications

Storage format of the hostnames by the applications is outside of the
scope of this protocol. 


2.2.4 Interface between applications and resolvers

This protocol does not specify the APIs that applications should use
to request the resolver to perform the DNS name resolution of the
internationalized hostnames. Instead it only specifies the format of
the hostnames specified in the input and output of such APIs.

The applications supporting non-ASCII characters in hostnames MUST
pass to the resolvers a hostname in ISO/IEC 10646 encoding. If the
response returned by the resolver to the application contains the
hostname, then the application should expect the hostname to be
encoded using ISO/IEC 10646.


2.2.5 Resolvers

Before sending the hostname in the query packet, the resolver MUST
prepare each name part as specified in [NAMEPREP]. After the name
preparation the resolver MUST convert the hostname to be encoded using
UTF-8 as specified in [RFC2044].
Names encoded in UTF-8 must not exceed the size limits clarified in
[RFC2181]. Character count is insufficient to determine size, since
some UTF-8 characters exceed one octet in length.




Expires November 2001                                         [Page 3]

INTERNET-DRAFT                  UTF-8 DNS                     May 2001


When resolver receives a response to the query from a DNS server, it
MUST convert all of the hostnames from UTF-8 encoded format to the
ISO/IEC 10646 encoding before passing these hostnames back to the
application. 


2.2.6 DNS servers

DNS servers authoritative for the records containing the hostnames
containing the characters not allowed by [RFC1123] MUST allow use of
the namepreped UTF-8 format to store and transmit those parts of the
hostnames.

According to existing standards, any binary string can be used in a
DNS name [RFC2181], but names must be compared with case-insensitivity
[RFC1035]. At the same time DNS protocol standard states that original
case SHOULD be preserved when possible as data is entered into the DNS
database. This requirement is modified as follows: a DNS server
authoritative for the internationalized hostnames MUST nameprep and
perform UTF-8 conversion on all names containing internationalized
characters in both record names and record data before storing these
hostnames and transmitting those names in any message. This new
requirement guarantees case-insensitive comparison of the
internationalized hostnames even by those DNS servers that do not
support this protocol.

DNS servers must compare names that contain UTF-8 characters
byte-for-byte, as opposed to using Unicode equivalency rules.


3. Interoperability Considerations

If user continues using ASCII-only characters in the hostnames, then
there is no need to upgrade any applications and/or resolvers.