RFC1506 - A Tutorial on Gatewaying between X.400 and Internet Mail
Network Working Group J. Houttuin
Request for Comments: 1506 RARE Secretariat
RARE Technical Report: 6 August 1993
A Tutorial on Gatewaying between X.400 and Internet Mail
Status of this Memo
This memo PRovides information for the Internet community. It does
not specify an Internet standard. Distribution of this memo is
unlimited.
IntrodUCtion
There are many ways in which X.400 and Internet (STD 11, RFC822)
mail systems can be interconnected. Addresses and service elements
can be mapped onto each other in different ways. From the early
available gateway implementations, one was not necessarily better
than another, but the sole fact that each handled the mappings in a
different way led to major interworking problems, especially when a
message (or address) crossed more than one gateway. The need for one
global standard on how to implement X.400 - Internet mail gatewaying
was satisfied by the Internet Request For Comments 1327, titled
"Mapping between X.400(1988)/ISO 10021 and RFC822."
This tutorial was produced especially to help new gateway managers
find their way into the complicated subject of mail gatewaying
according to RFC1327. The need for such a tutorial can be
illustrated by quoting the following discouraging paragraph from RFC
1327, chapter 1: "Warning: the remainder of this specification is
technically detailed. It will not make sense, except in the context
of RFC822 and X.400 (1988). Do not attempt to read this document
unless you are familiar with these specifications."
The introduction of this tutorial is general enough to be read not
only by gateway managers, but also by e-mail managers who are new to
gatewaying or to one of the two e-mail worlds in general. Parts of
this introduction can be skipped as needed.
For novice end-users, even this tutorial will be difficult to read.
They are encouraged to use the COSINE MHS pocket user guide [14]
instead.
To a certain extent, this document can also be used as a reference
guide to X.400 <-> RFC822 gatewaying. Wherever there is a lack of
detail in the tutorial, it will at least point to the corresponding
chapters in other documents. As such, it shields the RFC1327 novice
from too much detail.
Acknowledgements
This tutorial is heavily based on other documents, such as [2], [6],
[7], [8], and [11], from which large parts of text were reproduced
(slightly edited) by kind permission from the authors.
The author would like to thank the following persons for their
thorough reviews: Peter Cowen (Nexor), Urs Eppenberger (SWITCH), Erik
Huizer (SURFnet), Steve Kille (ISODE Consortium), Paul Klarenberg
(NetConsult), Felix Kugler (SWITCH), Sabine Luethi.
Disclaimer
This document is not everywhere exact and/or complete in describing
the involved standards. Irrelevant details are left out and some
concepts are simplified for the ease of understanding. For reference
purposes, always use the original documents.
Table of Contents
1. An overview of relevant standards ........................ 4
1.1. What is X.400 ? ...................................... 5
1.2. What is an RFC? ..................................... 8
1.3. What is RFC822 ? .................................... 9
1.4. What is RFC1327 ? ................................... 11
2. Service Elements ......................................... 12
3. Address mapping .......................................... 14
3.1. X.400 addresses ...................................... 15
3.1.1. Standard Attributes .............................. 15
3.1.2. Domain Defined Attributes ........................ 17
3.1.3. X.400 address notation ........................... 17
3.2. RFC822 addresses .................................... 19
3.3. RFC1327 address mapping ............................. 20
3.3.1. Default mapping .................................. 20
3.3.1.1. X.400 -> RFC822 ............................. 20
3.3.1.2. RFC822 -> X.400 ............................. 22
3.3.2. Exception mapping ................................ 23
3.3.2.1. PersonalName and localpart mapping ........... 25
3.3.2.2. X.400 domain and domainpart mapping .......... 26
3.3.2.2.1. X.400 -> RFC822 ......................... 27
3.3.2.2.2. RFC822 -> X.400 ......................... 28
3.4. Table co-ordination .................................. 31
3.5. Local additions ...................................... 31
3.6. Product specific formats ............................. 32
3.7. Guidelines for mapping rule definition ............... 34
4. Conclusion ............................................... 35
Appendix A. References ...................................... 36
Appendix B. Index (Only available in the Postscript version) 37
Appendix C. Abbreviations ................................... 37
Appendix D. How to access the MHS Co-ordination Server ...... 38
Security Considerations ..................................... 39
Author's Address ............................................ 39
1. An overview of relevant standards
This chapter describes the history, status, future, and contents of
the involved standards.
There is a major difference between mail systems used in the USA and
Europe. Mail systems originated mainly in the USA, where their
eXPlosive growth started as early as in the seventies. Different
company-specific mail systems were developed simultaneously, which,
of course, led to a high degree of incompatibility. The Advanced
Research Projects Agency (ARPA), which had to use machines of many
different manufacturers, triggered the development of the Internet
and the TCP/IP protocol suite, which was later accepted as a standard
by the US Department of Defense (DoD). The Internet mail format is
defined in STD 11, RFC822 and the protocol used for exchanging mail
is known as the simple mail transfer protocol (SMTP) [1]. Together
with UUCP and the BITNET protocol NJE, SMTP has become one of the
main de facto mail standards in the US.
Unfortunately, all these protocols were incompatible, which explains
the need to come to an acceptable global mail standard. CCITT and
ISO began working on a norm and their work converged in what is now
known as the X.400 Series Recommendations. One of the objectives was
to define a superset of the existing systems, allowing for easier
integration later on. Some typical positive features of X.400 are the
store-and-forward mechanism, the hierarchical address space and the
possibility of combining different types of body parts into one
message body.
In Europe, the mail system boom came later. Since there was not much
equipment in place yet, it made sense to use X.400 as much as
possible right from the beginning. A strong X.400 lobby existed,
especially in West-Germany (DFN). In the R&D world, mostly EAN was
used because it was the only affordable X.400 product at that time
(Source-code licenses were free for academic institutions).
At the moment, the two worlds of X.400 and SMTP are moving closer
together. For instance, the United States Department of Defense, one
of the early forces behind the Internet, has decided that future DoD
networking should be based on ISO standards, implying a migration
from SMTP to X.400. As an important example of harmonisation in the
other direction, X.400 users in Europe have a need to communicate
with the Internet. Due to the large traffic volume between the two
nets it is not enough interconnecting them with a single
international gateway. The load on such a gateway would be too
heavy. Direct access using local gateways is more feasible.
Although the expected success of X.400 has been a bit disappointing
(mainly because no good products were available), many still see the
future of e-mail systems in the context of this standard.
And regardless if in the long run X.400 will or will not take over
the world of e-mail systems, SMTP cannot be neglected over the next
ten years. Especially the simple installation procedures and the high
degree of connectivity will contribute to a growing number of RFC822
installations in Europe and world-wide in the near future.
1.1. What is X.400 ?
In October 1984, the Plenary Assembly of the CCITT accepted a
standard to facilitate international message exchange between
subscribers to computer based store-and-forward message services.
This standard is known as the CCITT X.400 series recommendations
([16], from now on called X.400(84)) and happens to be the first
CCITT recommendation for a network application. It should be noted
that X.400(84) is based on work done in the IFIP Working Group 6.5,
and that ISO at the same time was proceeding towards a compatible
document. However, the standardisation efforts of CCITT and ISO did
not converge in time (not until the 1988 version), to allow the
publication of a common text.
X.400(84) triggered the development of software implementing (parts
of) the standard in the laboratories of almost all major computer
vendors and many software houses. Similarly, public carriers in many
countries started to plan X.400(84) based message systems that would
be offered to the users as value added services. Early
implementations appeared shortly after first drafts of the standard
were published and a considerable number of commercial systems are
available nowadays.
X.400(84) describes a functional model for a Message Handling System
(MHS) and associates services and protocols. The model illustrated in
Figure 1.1. defines the components of a distributed messaging system.
Users in the MHS environment are provided with the capability of
sending and receiving messages. Users in the context of an MHS may be
humans or application processes. The User Agent (UA) is a process
that makes the services of the MTS available to the user. A UA may be
implemented as a computer program that provides utilities to create,
send, receive and perhaps archive messages. Each UA, and thus each
user, is identified by a name (each user has its own UA).
-----------------------------------------------------------------
user user Message Handling Environment
----------------------------------------------------------
Message Handling System
---- ----
UA UA
---- ----
-------------------------------------------------
Message Transfer System
---- ----- -----
user--UA----MTA MTA
---- ----- -----
\ /
\ /
\ /
\ /
\ /
---- -----
user--UA-----------MTA
---- -----
-------------------------------------------------
----------------------------------------------------------
-----------------------------------------------------------------
Fig. 1.1. X.400 functional model
The Message Transfer system (MTS) transfers messages from an
originating UA to a recipient UA. As implied by the Figure 1.1, data
sent from UA to UA may be stored temporarily in several intermediate
Message Transfer Agents (MTA), i.e., a store-and- forward mechanism
is being used. An MTA forwards received messages to a next MTA or to
the recipient UA.
X.400(84) divides layer 7 of the OSI Reference Model into 2
sublayers, the User Agent Layer (UAL) and the Message Transfer Layer
(MTL) as shown in the Figure 1.2.
The MTL is involved in the transport of messages from UA to UA, using
one or several MTAs as intermediaries. By consequence, routing issues
are entirely dealt with in the MTL. The MTL in fact corresponds to
the postal service that forwards letters consisting of an envelope
and a content. Two protocols, P1 and P3, are used between the MTL
entities (MTA Entity (MTAE), and Submission and Delivery Entity
(SDE)) to reliably transport messages. The UAL embodies peer UA
Entities (UAE), which interpret the content of a message and offer
specific services to the application process. Depending on the
application to be supported on top of the MTL, one of several end-
to-end protocols (Pc) is used between UAEs. For electronic mail,
X.400(84) defines the protocol P2 as part of the InterPersonal
Messaging Service (IPMS). Conceivably other UAL protocols may be
defined, e.g., a protocol to support the exchange of electronic
business documents.
--------------------------------------------------------------
----- -----
UA layer UAE<----- P2, Pc ----------->UAE
----- -----
--------------------------------------------------------------
------ ------ -----
MTA layer MTAE<-- P1 -->MTAE<-- P3-->SDE
------ ------ -----
--------------------------------------------------------------
xxxE = xxx Entity ; SDE = Submission & Delivery Entity
--------------------------------------------------------------
Fig. 1.2. X.400 Protocols
The structure of an InterPersonal Message (IPM) can be visualised as
in Figure 1.3. (Note that the envelope is not a part of the IPM; it
is generated by the MTL).
Forwarded
Message IP-message
- ---------- --- ---------- -
message- envelope / PDI
content IPM ---------- / ----------
- - ---------- / ----------
IPM- heading / heading
body ---------- / ----------
- ----------/ ----------
bodypart bodypart
----------\ ----------
---------- \ ----------
bodypart \ bodypart
---------- \ ----------
. \
. \
---------- \ ----------
bodypart \ bodypart
- - - - ---------- - ---------- -
(PDI = Previous Delivery Info.)
Fig. 1.3. X.400 message structure
An IPM heading contains information that is specific for an
interpersonal message like 'originator', 'subject', etc. Each
bodypart can contain one information type, text, voice or as a
special case, a forwarded message. A forwarded message consists of
the original message together with Previous Delivery Information
(PDI), which is drawn from the original delivery envelope.
Early experience with X.400(84) showed that the standard had various
shortcomings. Therefore CCITT, in parallel with ISO, corrected and
extended the specification during its 1984 to 1988 study period and
produced a revised standard [17], which was accepted at the 1988
CCITT Plenary Meeting [10]. Amongst others, X.400(88) differs from
X.400(84) in that it defines a Message Store (MS), which can be seen
as a kind of database for messages. An MS enables the end-user to run
a UA locally, e.g., on a PC, whilst the messages are stored in the
MS, which is co-located with the MTA. The MTA can thus always deliver
incoming messages to the MS instead of to the UA. The MS can even
automatically file incoming messages according to certain criteria.
Other enhancements in the 88 version concern security and
distribution lists.
1.2. What is an RFC?
The Internet, a loosely-organised international collaboration of
autonomous, interconnected networks, supports host-to-host
communication through voluntary adherence to open protocols and
procedures defined by Internet Standards. There are also many
isolated internets, i.e., sets of interconnected networks, that are
not connected to the Internet but use the Internet Standards. The
architecture and technical specifications of the Internet are the
result of numerous research and development activities conducted over
a period of two decades, performed by the network R&D community, by
service and equipment vendors, and by government agencies around the
world.
In general, an Internet Standard is a specification that is stable
and well-understood, is technically competent, has multiple,
independent, and interOperable implementations with operational
experience, enjoys significant public support, and is recognisably
useful in some or all parts of the Internet.
The principal set of Internet Standards is commonly known as the
"TCP/IP protocol suite". As the Internet evolves, new protocols and
services, in particular those for Open Systems Interconnection (OSI),
have been and will be deployed in traditional TCP/IP environments,
leading to an Internet that supports multiple protocol suites.
The following organisations are involved in setting Internet
standards.
Internet standardisation is an organised activity of the Internet
Society (ISOC). The ISOC is a professional society that is concerned
with the growth and evolution of the world-wide Internet, with the
way in which the Internet is and can be used, and with the social,
political, and technical issues that arise as a result.
The Internet Engineering Task Force (IETF) is the primary body
developing new Internet Standard specifications. The IETF is composed
of many Working Groups, which are organised into areas, each of which
is co-ordinated by one or more Area Directors.
The Internet Engineering Steering Group (IESG) is responsible for
technical management of IETF activities and the approval of Internet
standards specifications, using well-defined rules. The IESG is
composed of the IETF Area Directors, some at-large members, and the
chairperson of the IESG/IETF.
The Internet Architecture Board (IAB) has been chartered by the
Internet Society Board of Trustees to provide quality control and
process appeals for the standards process, as well as external
technical liaison, organizational oversight, and long-term
architectural planning and research.
Any individual or group (e.g., an IETF or RARE working group) can
submit a document as a so-called Internet Draft. After the document
is proven stable, the IESG may turn the Internet-Draft into a
"Requests For Comments" (RFC). RFCs cover a wide range of topics,
from early discussion of new research concepts to status memos about
the Internet. All Internet Standards (STDs) are published as RFCs,
but not all RFCs specify standards. Another sub-series of the RFCs
are the RARE Technical Reports (RTRs).
As an example, this tutorial also started out as an Internet-Draft.
After almost one year of discussions and revisions it was approved by
the IESG as an Informational RFC.
Once a document is assigned an RFCnumber and published, that RFCis
never revised or re-issued with the same number. Instead, a revision
will lead to the document being re-issued with a higher number
indicating that an older one is obsoleted.
1.3. What is RFC822 ?
STD 11, RFC822 defines a standard for the format of Internet text
messages. Messages consist of lines of text. No special provisions
are made for encoding drawings, facsimile, speech, or structured
text. No significant consideration has been given to questions of
data compression or to transmission and storage efficiency, and the
standard tends to be free with the number of bits consumed. For
example, field names are specified as free text, rather than special
terse codes.
A general "memo" framework is used. That is, a message consists of
some information in a rigid format (the 'headers'), followed by the
main part of the message (the 'body'), with a format that is not
specified in STD 11, RFC822. It does define the syntax of several
fields of the headers section; some of these fields must be included
in all messages.
STD 11, RFC822 is used in conjunction with a number of different
message transfer protocol environments (822-MTSs).
- SMTP Networks: On the Internet and other TCP/IP networks,
STD 11, RFC822 is used in conjunction with two other
standards: STD 10, RFC821, also known as Simple Mail
Transfer Protocol (SMTP) [1], and RFCs 1034 and 1035
which specify the Domain Name System [3].
- UUCP Networks: UUCP is the UNIX to UNIX CoPy protocol, which
is usually used over dialup telephone networks to provide a
simple message transfer mechanism.
- BITNET: Some parts of Bitnet and related networks use STD
11, RFC822 related protocols, with EBCDIC encoding.
- JNT Mail Networks: A number of X.25 networks, particularly
those associated with the UK Academic Community, use the JNT
(Joint Network Team) Mail Protocol, also known as Greybook.
STD 11, RFC822 is based on the assumption that there is an
underlying service, which in RFC1327 is called the 822-MTS service.
The 822-MTS service provides three basic functions:
1. Identification of a list of recipients.
2. Identification of an error return address.
3. Transfer of an RFC822 message.
It is possible to achieve 2) within the RFC822 header. Some 822-
MTS protocols, in particular SMTP, can provide additional
functionality, but as these are neither mandatory in SMTP, nor
available in other 822-MTS protocols, they are not considered here.
Details of aspects specific to two 822-MTS protocols are given in
Appendices B and C of RFC1327. An RFC822 message consists of a
header, and content which is uninterpreted ASCII text. The header is
divided into fields, which are the protocol elements. Most of these
fields are analogous to P2 heading fields, although some are
analogous to MTS Service Elements.
1.4. What is RFC1327 ?
There is a large community using STD 11, RFC822 based protocols for
mail services, who will wish to communicate with users of the
InterPersonal Messaging Service (IPMS) provided by X.400 systems, and
the other way around. This will also be a requirement in cases where
RFC822 communities intend to make a transition to use X.400 (or the
other way around, which also happens), as conversion will be needed
to ensure a smooth service transition.
The basic function of a mail gateway can be described as follows:
receive a mail from one mail world, translate it into the formats of
the other mail world and send it out again using the routing rules
and protocols of that other world.
Especially if a message crosses more than one gateway, it is
important that all gateways have the same understanding of how things
should be mapped. A simple example of what could go wrong otherwise
is the following: A sends a message to B through a gateway and B's
reply to A is being routed through another gateway.
If the two gateways don't use the same mappings, it can be expected
that the From and To addresses in the original mail and in the answer
don't match, which is, to say the least, very confusing for the end-
users (consider what happens if automated processes communicate via
mail). More serious things can happen to addresses if a message
crosses more than one gateway on its way from the originator to the
recipient. As a real-life example, consider receiving a message from:
Mary Plork <MMP_+a_ARG_+lMary_Plork+r%MHS+d_A0CD8A2B01F54FDC-
A0CB9A2B03F53FDC%[email protected]Gmail.com>
This is not what you would call user-friendly addressing.... RFC1327
describes a set of mappings that will enable a more transparent
interworking between systems operating X.400 (both 84 and 88) and
systems using RFC822, or protocols derived from STD 11, RFC822.
RFC1327 describes all mappings in term of X.400(88). It defines how
these mappings should be applied to X.400(84) systems in its Appendix
G.
Some Words about the history of RFC1327: It started out in June
1986, when RFC987 defined for X.400(84) what RFC1327 defines for
X.400(84 and 88). RFC1026 specified a number of additions and
corrections to RFC987. In December 1989, RFC1138, which had a very
short lifetime, was the first one to deal with X.400(88). It was
obsoleted by RFC1148 in March 1990. Finally, in May 1992, RFC1327
obsoleted all of its ancestors.
2. Service Elements
Both RFC822 and X.400 messages consist of certain service elements
(such as 'originator' and 'subject'). As long as a message stays
within its own world, the behaviour of such service elements is well
defined. An important goal for a gateway is to maintain the highest
possible service level when a message crosses the boundary between
the two mail worlds.
When a user originates a message, a number of services are available.
RFC1327 describes, for each service elements, to what extent it is
supported for a recipient accessed through a gateway. There are
three levels of support:
- Supported: Some of the mappings are quite straight-forward,
such as '822.Subject:' <-> 'IPMS.Subject'.
- Not supported: There may be a complete mismatch: certain
service elements exist only in one of the two worlds (e.g.,
interpersonal notifications).
- Partially supported: When similar service elements exist in
both worlds, but with slightly different interpretations,
some tricks may be needed to provide the service over the
gateway border.
Apart from mapping between the service elements, a gateway must also
map the types and values assigned to these service elements. Again,
this may in certain cases be very simple, e.g., 'IA5 -> ASCII'. The
most complicated example is mapping address spaces. The problem is
that address spaces are not something static that can be defined
within RFC1327. Address spaces change continuously, and they are
defined by certain addressing authorities, which are not always
parallel in the RFC822 and the X.400 world. A valid mapping between
two addresses assumes however that there is 'administrative
equivalence' between the two domains in which the addresses exist
(see also [13]).
The following basic mappings are defined in RFC1327. When going from
RFC822 to X.400, an RFC822 message and the associated 822- MTS
information is always mapped into an IPM (MTA, MTS, and IPMS
Services). Going from X.400 to RFC822, an RFC822 message and the
associated 822-MTS information may be derived from:
- A Report (MTA, and MTS Services)
- An InterPersonal Notification (IPN) (MTA, MTS, and IPMS
services)
- An InterPersonal Message (IPM) (MTA, MTS, and IPMS services)
Probes (MTA Service) have no equivalent in STD 10, RFC821 or STD 11,
RFC822 and are thus handled by the gateway. The gateway's Probe
confirmation should be interpreted as if the gateway were the final
MTA to which the Probe was sent. Optionally, if the gateway uses RFC
821 as an 822-MTS, it may use the results of the 'VRFY' command to
test whether it would be able to deliver (or forward) mail to the
mailbox under probe.
MTS Messages containing Content Types other than those defined by the
IPMS are not mapped by the gateway, and should be rejected at the
gateway.
Some basic examples of mappings between service elements are listed
below.
Service elements:
RFC822 X.400
------------------------------------------------
Reply-To: IPMS.Heading.reply-recipients
Subject: IPMS.Heading.subject
In-Reply-To: IPMS.Heading.replied-to-ipm
References: IPMS.Heading.related-IPMs
To: IPMS.Heading.primary-recipients
Cc: IPMS.Heading.copy-recipients
Service element types:
RFC822 X.400
------------------------------------------------
ASCII PrintableString
Boolean Boolean
Service element values:
RFC822 X.400
------------------------------------------------
oh_dear oh(u)dear
False 00000000
There are some mappings between service elements that are rather
tricky and important enough to mention in this tutorial. These are
the mappings of origination-related headers and some envelope fields:
RFC822 -> X.400:
- If Sender: is present, Sender: is mapped to
IPMS.Heading.originator, and From: is mapped to
IPMS.Heading.authorizing-users. If not, From: is mapped to
IPMS.Heading.originator.
X.400 -> RFC822
- If IPMS.Heading.authorizing-users is present,
IPMS.Heading.originator is mapped to Sender:, and
IPMS.Heading.authorizing-users is mapped to From: . If not,
IPMS.Heading.originator is mapped to From:.
Envelope attributes
- RFC1327 doesn't define how to map the MTS.OriginatorName and
the MTS.RecipientName (often referred to as the P1.originator
and P1.recipient), since this depends on which underlying 822-
MTS is used. In the very common case that RFC821 (SMTP) is
used for this purpose, the mapping is normally as follows:
MTS.Originator-name <-> MAIL FROM:
MTS.Recipient-name <-> RCPT TO:
For more details, refer to RFC1327, chapters 2.2 and 2.3.
3. Address mapping
As address mapping is often considered the most complicated part of
mapping between service element values, this subject is given a
separate chapter in this tutorial.
Both RFC822 and X.400 have their own specific address formats. RFC
822 addresses are text strings (e.g., "[email protected]"), whereas X.400
addresses are binary encoded sets of attributes with values. Such
binary addresses can be made readable for a human user by a number of
notations; for instance:
C=zz
ADMD=ade
PRMD=fhbo
O=a bank
S=plork
G=mary
The rest of this chapter deals with addressing issues and mappings
between the two address forms in more detail.
3.1. X.400 addresses
As already stated above, an X.400 address is modelled as a set of
attributes. Some of these attributes are mandatory, others are
optional. Each attribute has a type and a value, e.g., the Surname
attribute has type IA5text, and an instance of this attribute could
have the value 'Kille'. Attributes are divided into Standard
Attributes (SAs) and Domain Defined Attributes (DDAs).
X.400 defines four basic forms of addresses ([17], 18.5), of which
the 'Mnemonic O/R Address' is the form that is most used, and is the
only form that is dealt with in this tutorial. This is roughly the
same address format as what in the 84 version was known as 'O/R
names: form 1, variant 1' ([16] 3.3.2).
3.1.1. Standard Attributes
Standard Attributes (SAs) are attributes that all X.400 installations
are supposed to 'understand' (i.e., use for routing), for example:
'country name', 'given name' or 'organizational unit'. The most
commonly used SAs in X.400(84) are:
surName (S)
givenName (G)
initials (I*) (Zero or more)
generationQualifier (GQ)
OrganizationalUnits (OU1 OU2 OU3 OU4)
OrganizationName (O)
PrivateDomainName (PRMD)
AdministrationDomainName (ADMD)
CountryName (C)
The combination of S, G, I* and GQ is often referred to as the
PersonalName (PN).
Although there is no hierarchy (of addressing authorities) defined by
the standards, the following hierarchy is considered natural:
PersonalName < OU4 < OU3 < OU2 < OU1 < O < P < A < C
In addition to the SAs listed above, X.400(88) defines some extra
attributes, the most important of which is
Common Name (CN)
CN can be used instead of or even together with PN. The problem in
X.400(84) was that PN (S G I* GQ) was well suited to represent
persons, but not roles and abstract objects, such as distribution
lists. Even though postmaster clearly is a role, not someone's real
surname, it is quite usual in X.400(84) to address a postmaster with
S=postmaster. In X.400(88), the same postmaster would be addressed
with CN=postmaster .
The attributes C and ADMD are mandatory (i.e., they must be present),
and may not be empty. At least one of the attributes PRMD, O, OU, PN
and CN must be present.
PRMD and ADMD are often felt to be routing attributes that don't
really belong in addresses. As an example of how such address
attributes can be used for the purpose of routing, consider two
special values for ADMD:
- ADMD=0; (zero) should be interpreted as 'the PRMD in this
address is not connected to any ADMD'
- ADMD= ; (single SPACE) should be interpreted as 'the PRMD in
this address is reachable via any ADMD in this country'. It
is expected that ISO will express this 'any' value by means
of a missing ADMD attribute in future versions of MOTIS.
This representation can uniquely identify the meaning 'any',
as a missing or empty ADMD field as such is not allowed.
Addresses are defined in X.400 using the Abstract Syntax Notation One
(ASN.1). X.409 defines how definitions in ASN.1 should be encoded
into binary format. Note that the meaning, and thus the ASN.1
encoding, of a missing attribute is not the same as that of an empty
attribute. In addressing, this difference is often represented as
follows:
- PRMD=; means that this attribute is present in the address,
but its value is empty. Since this is not very useful, it's
hardly ever used. The only examples the author knows of
were caused by mail managers who should have had this
tutorial before they started defining their addresses :-)
- [email protected]; means that this attribute is not present in the
address. {NB. This is only necessary if an address notation
(see 3.1.3) requires that every single attribute in the
hierarchy is somehow listed. Otherwise, a missing attribute
can of course be represented by simply not mentioning it.
This means that this syntax is mostly used in mapping rules,
not by end users.}
Addresses that only contain SAs are often referred to as Standard
Attribute Addresses (SAAs).
3.1.2. Domain Defined Attributes
Domain Defined Attributes (DDAs) can be used in addition to Standard
Attributes. An instance of a DDA consists of a type and a value. DDAs
are meant to have a meaning only within a certain context (originally
this was supposed to be the context of a certain management domain,
hence the name DDA), such as a company context.
As an example, a company might want to define a DDA for describing
internal telephone numbers: DDA type=phone value=9571.
A bit tricky is the use of DDAs to encode service element types or
values that are only available on one side of a service gateway. The
most important examples of such usage are defined in:
RFC1327 (e.g., DDA type=RFC-822 value=u(u)ser(a)isode.com)
RFC1328 (e.g., DDA type=CommonName value=mhs-discussion-list)
Addresses that contain both SAs and DDAs are often referred to as DDA
addresses.
3.1.3. X.400 address notation
X.400 only prescribes the binary encoding of addresses, it doesn't
standardise how such addresses should be written on paper or what
they should look like in a user interface on a computer screen.
There exist a number of recommendations for X.400 address
representation though.
- JTC proposed an annex to CCITT Rec. F.401 and ISO/IEC 10021-2,
called 'Representation of O/R addresses for human usage'. According
to this proposal, an X.400 address would look as follows:
G=jo; S=plork; O=a bank; OU1=owe; OU2=you; P=fhbo; A=ade; C=zz
Note that in this format, the order of O and the OUs is exactly
the opposite of what one would expect intuitively (the attribute
hierarchy is increasing from left to right, except for the O and
OUs, where it's right to left. The reasoning behind this is that
this sequence is following the example of a postal address). This
proposal has been added (as a recommendation) to the 1992 version
of the standards.
- Following what was originally used in the DFN-EAN software, most
EAN versions today use an address representation similar to the JTC
proposal, with a few differences:
- natural ordering for O and OUs
- no numbering of OUs.
- allows writing ADMD and PRMD instead of A and P
The address in the example above could, in EAN, be represented as:
G=jo; S=plork; OU=you; OU=owe; O=a bank; PRMD=fhbo; ADMD=ade; C=zz
This DFN-EAN format is still often referred to as _the_ 'readable
format'.
- The RARE Working Group on Mail and Messaging, WG-MSG, has made a
recommendation that is very similar to the DFN-EAN format, but with
the hierarchy reversed. Further, ADMD and PRMD are used instead of
A and P. This results in the address above being represented as:
C=zz; ADMD=ade; PRMD=fhbo; O=a bank; OU=owe; OU=you; S=plork; G=jo
This format is recognised by most versions of the EAN software. In
the R&D community, this is one of the most popular address
representations for business cards, letter heads, etc. It is also
the format that will be used for the examples in this tutorial.
(NB. The syntax used here for describing DDAs is as follows:
DD.'type'='value', e.g., DD.phone=9571)
- RFC1327 defines a slash separated address representation:
/G=jo/S=plork/OU=you/OU=owe/O=a bank/P=fhbo/A=ade/C=zz/
Not only is this format used by the PP software, it is also
widespread for business cards and letter heads in the R&D
community.
- RFC1327 finally defines yet another format for X.400 _domains_
(not for human users):
OU$you.OU$owe.O$a bank.P$fhbo.A$ade.C$zz
The main advantage of this format is that it is better machine-
parseble than the others, which also immediately implies its main
disadvantage: it is barely readable for humans. Every attribute
within the hierarchy should be listed, thus a missing attribute
must be represented by the '@' sign
(e.g., $a [email protected]$ade.C$zz).
- Paul-Andre Pays (INRIA) has proposed a format that combines the
readability of the JTC format with the parsebility of the RFC1327
domain format. Although a number of operational tools within the GO-
MHS community are already based on (variants of) this proposal, its
future is still uncertain.
3.2. RFC822 addresses
An RFC822 address is an ASCII string of the following form:
[email protected]
"domainpart" is sub-divided into
domainpart = sdom(n).sdom(n-1)....sdom(2).sdom(1).dom
"sdom" stands for "subdomain", "dom" stands for "top-level-domain".
"localpart" ;is normally a login name, and thus typically is a
surname or an abbreviation for this. It can also designate a local
distribution list.
The hierarchy (of addressing authorities) in an RFC822 address is
as follows:
localpart < sdom(n) < sdom(n-1) <...< dom
Some virtual real-life examples:
[email protected]
[email protected]
[email protected]
In the above examples, 'nl', 'dk', and 'edu' are valid,
registered, top level domains. Note that some networks that have
their own addressing schemes are also reachable by way of 'RFC
822-like' addressing. Consider the following addresses:
oops!user (a UUCP address)
[email protected] (a BITNET address)
These addresses can be expressed in RFC822 format:
[email protected]
[email protected]
Note that the domains '.uucp' and '.bitnet' have no registered
Internet routing. Such addresses must always be routed to a gateway
(how this is done is outside the scope of this tutorial).
As for mapping such addresses to X.400, there is no direct mapping
defined between X.400 on the one hand and UUCP and BITNET on the
other, so they are normally mapped to RFC822 style first, and then
to X.400 if needed.
3.3. RFC1327 address mapping
Despite the difference in address formats, the address spaces defined
by RFC822 and X.400 are quite similar. The most important parallels
are:
- both address spaces are hierarchical
- top level domains and country codes are often the same
- localparts and surnames are often the same
This similarity can of course be exploited in address mapping
algorithms. This is also done in RFC1327 (NB only in the exception
mapping algorithm. See chapter 3.3.2).
Note that the actual mapping algorithm is much more complicated than
shown below. For details, see RFC1327, chapter 4.
3.3.1. Default mapping
The default RFC1327 address mapping can be visualised as a function
with input and output parameters:
address information of the gateway performing the mapping
v
+-----------------+
RFC822 address <---> address mapping <---> X.400 address
+-----------------+
I.e., to map an address from X.400 to RFC822 or vice versa, the only
extra input needed is the address information of the local gateway.
3.3.1.1. X.400 -> RFC822
There are two kinds of default address mapping from X.400 to RFC822:
one to map a real X.400 address to RFC822, and another to decode an
RFC822 address that was mapped to X.400 (i.e., to reverse the
default RFC822 -> X.400 mapping).
To map a real X.400 address to RFC822, the slash separated notation
of the X.400 address (see chapter 3.1.) is mapped to 'localpart', and
the local RFC822 domain of the gateway that performs the mapping is
used as the domain part. As an example, the gateway 'gw.switch.ch'
would perform the following mappings:
C=zz; ADMD=ade; PRMD=fhbo; O=tlec; S=plork; ->
/C=zz/ADMD=ade/PRMD=fhbo/O=tlec/S=plork/@gw.switch.ch
C=zz; ADMD=ade; PRMD=fhbo; O=a bank; S=plork->
"/C=zz/ADMD=ade/PRMD=fhbo/O=a bank/S=plork/"@gw.switch.ch
The quotes in the second example are mandatory if the X.400 address
contains spaces, otherwise the syntax rules for the RFC822 localpart
would be violated.
This default mapping algorithm is generally referred to as 'left-
hand-side encoding'.
To reverse the default RFC822 -> X.400 mapping (see chapter
3.3.1.2): if the X.400 address contains a DDA of the type RFC-822,
the SAs can be discarded, and the value of this DDA is the desired
RFC822 address (NB. Some characters in the DDA value must be decoded
first. See chapter 3.3.1.2.). For example, the gateway
DD.RFC-822=bush(a)dole.us; C=nl; ADMD=tlec; PRMD=GW
->
[email protected]
3.3.1.2. RFC822 -> X.400
There are also two kinds of default address mapping from RFC822 to
X.400: one to map a real RFC822 address to X.400, and another to
decode an X.400 address that was mapped to RFC822 (i.e., to reverse
the default X.400 -> RFC822 mapping).
To map a real RFC822 address to X.400, the RFC822 address is
encoded in a DDA of type RFC-822 , and the SAs of the local gateway
performing the mapping are added to form the complete X.400 address.
This mapping is generally referred to as 'DDA mapping'. As an
example, the gateway 'C=nl; ADMD=tlec; PRMD=GW' would perform the
following mapping:
[email protected] ->
DD.RFC-822=bush(a)dole.us; C=nl; ADMD=tlec; PRMD=GW
As for the encoding/decoding of RFC822 addresses in DDAs, it is
noted that RFC822 addresses may contain characters (@ ! % etc.) that
cannot directly be represented in a DDA. DDAs are of the restricted
character set type 'PrintableString', which is a subset of IA5
(=ASCII). Characters not in this set need a special encoding. Some
examples (For details, refer to RFC1327, chapter 3.4.):
100%[email protected] -> DD.RFC-822;=100(p)name(a)address
[email protected] -> DD.RFC-822;=u(u)ser(b)name(a)address
To decode an X.400 address that was mapped to RFC822: if the RFC822
address has a slash separated representation of a complete X.400
mnemonic O/R address in its localpart, that address is the result of
the mapping. As an example, the gateway 'gw.switch.ch' would perform
the following mapping:
/C=zz/ADMD=ade/PRMD=fhbo/O=tlec/S=plork/G=mary/@gw.switch.ch
->
C=zz; ADMD=ade; PRMD=fhbo; O=tlec; S=plork; G=mary
3.3.2. Exception mapping according to mapping tables
Chapter 3.3.1. showed that it is theoretically possible to use RFC
1327 with default mapping only. Although this provides a very simple,
straightforward way to map addresses, there are some very good
reasons not to use RFC1327 this way:
- RFC822 users are used to writing simple addresses of the
form '[email protected]'. They often consider X.400
addresses, and thus also the left-hand-side encoded
equivalents, as unnecessarily long and complicated. They
would rather be able to address an X.400 user as if she had a
'normal' RFC822 address. For example, take the mapping
C=zz; ADMD=ade; PRMD=fhbo; O=tlec; S=plork; ->
/C=zz/ADMD=ade/PRMD=fhbo/O=tlec/S=plork/@gw.switch.ch
from chapter 3.3.1.1. RFC822 users would find it much more
'natural' if this address could be expressed in RFC822 as:
[email protected]
- X.400 users are used to using X.400 addresses with SAs only.
They often consider DDA addresses as complicated, especially
if they have to encode the special characters, @ % ! etc,
manually. They would rather be able to address an RFC822
user as if he had a 'normal' X.400 address. For example, take
the mapping
[email protected]
->
DD.RFC-822=bush(a)dole.us;
C=nl; ADMD= ; PRMD=tlec; O=gateway
from chapter 3.3.1.2. X.400 users would find it much more
'natural' if this address could be expressed in X.400 as:
C=us; ADMD=dole; S=bush
- Many organisations are using both RFC822 and X.400
internally, and still want all their users to have a simple,
unique address in both mail worlds. Note that in the default
mapping, the mapped form of an address completely depends on
which gateway performed the mapping. This also results in a
complication of a more technical nature:
- The tricky 'third party problem'. This problem need not
necessarily be understood to read the rest of this chapter.
If it looks too complicated, please feel free to skip it
until you are more familiar with the basics.
The third party problem is a routing problem caused by
mapping. As an example for DDA mappings (the example holds
just as well for left-hand-side encoding), consider the
following situation (see Fig. 3.1.): RFC822 user X in
country A sends a message to two recipients: RFC822 user Y,
and X.400 user Z, both in country B:
From: [email protected]
To: [email protected] ,
/C=B/.../S=Z/@GW.A
Since the gateway in country A maps all addresses in the
message, Z will see both X's and Y's address as DDA-encoded
RFC822 addresses, with the SAs of the gateway in country A:
From: DD.RFC-822=X(a)A; C=A;....;O=GW
To: DD.RFC-822=Y(a)B; C=A;....;O=GW ,
C=B;...;S=Z
------------ ---------
X: RFC822<------->gateway
------------ ---------
A ^
\
\---------------------------------------------
/---------------------------------------------
/
B v
-----------
Z: X.400
-----------
.
.
.
.
.
v v
------------ ---------
Y: RFC822<........gateway
------------ ---------
Fig. 3.1 The third party problem
Now if Z wants to 'group reply' to both X and Y, his reply to Y
will be routed over the gateway in country A, even though Y is
located in the same country:
From: C=B;...;S=Z
To: DD.RFC-822=Y(a)B; C=A;....;O=GW ,
DD.RFC-822=X(a)A; C=A;....;O=GW
The best way to travel for a message from Z to Y would of
course have been over the gateway in country B:
From: C=B;...;S=Z
To: DD.RFC-822=Y(a)B; C=B;....;O=GW ,
DD.RFC-822=X(a)A; C=A;....;O=GW
The third party problem is caused by the fact that routing
information is mapped into addresses.
Ideally, the third party problem shouldn't exist. After all,
address mapping affects addresses, and an address is not a
route.... The reality is different however. For instance, very
few X.400 products are capable to route messages on the
contents of a DDA (actually, only RFC1327 gateways will be
able to interpret this type of DDA, and who says that the reply
will pass a local gateway on its route back?). Similar
limitations hold for the other direction: an RFC822 based
mailer is not even allowed (see [5]) to make routing decisions
of the content of a left-hand-side encoded X.400 address if the
domain part is not its own. So in practice, addressing and
(thus also mapping) will very well affect routing.
To make mapping between addresses more user friendly, and to avoid
the problems shown above, RFC1327 allows for overruling the default
left-hand-side encoding and DDA mapping algorithms. This is done by
specifying associations (mapping rules) between certain domainparts
and X.400 domains. An X.400 domain (for our purposes; CCITT has a
narrower definition...) consists of the domain-related SAs of a
Mnemonic O/R address (i.e., all SAs except PN and CN). The idea is to
use the similarities between both address spaces, and directly map
similar address parts onto each other. If, for the domain in the
address to be mapped, an explicit mapping rule can be found, the
mapping is performed between:
localpart <-> PersonalName
domainpart <-> X.400 domain
The address information of the gateway is only used as an input
parameter if no mapping rule can be found, i.e., if the address
mapping must fall back to its default algorithm.
The complete mapping function can thus be visualised as follows:
address information of the gateway performing the mapping
v
+-----------------+
RFC822 address <---> address mapping <---> X.400 address
+-----------------+
^
domain associations (mapping rules)
3.3.2.1. PersonalName and localpart mapping
Since the mapping between these address parts is independent of the
mapping rules that are used, and because it follows a simple, two-
way algorithmic approach, this subject is discussed in a separate
sub-chapter first.
The X.400 PersonalName consists of givenName, initials, and surName.
RFC1327 assumes that generationQualifier is not used.
To map a localpart to an X.400 PN, the localpart is scanned for dots,
which are considered delimiters between the components of PN, and
also between single initials. In order not to put too much detail in
this tutorial, only a few examples are shown here. For the detailed
algorithm, see RFC1327, chapter 4.2.1.
Marshall.Rose <-> G=Marshall;S=Rose
M.T.Rose <-> I=MT;S=Rose
Marshall.M.T.Rose <-> G=Marshall;I=MT;S=Rose
To map an X.400 PN to an RFC822 localpart, take the non-empty PN
attributes, put them into their hierarchical order (G I* S), and
connect them with periods.
Some exceptions are caused by the fact that left-hand-side encoding
can also be mixed with exception mapping. This is shown in more
detail in the following sub-chapters.
3.3.2.2. X.400 domain and domainpart mapping
A mapping rule associates two domains: an X.400 domain and an RFC822
domain. The X.400 domain is written in the RFC1327 domain notation
(See 3.1.3.), so that both domains have the same hierarchical order.
The domains are written on one line, separated by a '#' sign. For
instance:
arcom.ch#ADMD$arcom.C$ch#
PRMD$tlec.ADMD$ade.C$nl#tlec.nl#
A mapping rule must at least contain a top level domain and a country
code. If an address must be mapped, a mapping rule with the longest
domain match is sought. The associated domain in the mapping rule is
used as the domain of the mapped address. The remaining domains are
mapped one by one following the natural hierarchy. Concrete examples
are shown in the following subchapters.
3.3.2.2.1. X.400 -> RFC822
As an example, assume the following mapping rule is defined:
PRMD$tlec.ADMD$ade.C$nl#tlec.nl#
Then the address C=nl; ADMD=ade; PRMD=tlec; O=you; OU=owe; S=plork
S OU O PRMD ADMD Country
plork owe you tlec ade nl
would be mapped as follows. The Surname 'plork' is mapped to the
localpart 'plork', see chapter 3.3.2.1. The domain
localpart
sdom3
sdom2
sdom1
top-level-domain
[email protected] tlec.nl
The remaining SAs (O and one OU) are mapped one by one following the
natural hierarchy: O is mapped to sdom2, OU is mapped to sdom3:
localpart
sdom3
sdom2
sdom1
top-level-domain
[email protected]
Thus the mapped address is:
[email protected]
The table containing the listing of all such mapping rules, which is
distributed to all gateways world-wide, is normally referred to as
'mapping table 1'. Other commonly used filenames (also depending on
which software your are using) are:
'or2rfc'
'mapping 1'
'map1'
'table 1'
'X2R'
As already announced, there is an exceptional case were localpart and
PN are not directly mapped onto each other: sometimes it is necessary
to use the localpart for other purposes. If the X.400 address
contains attributes that would not allow for the simple mapping:
localpart <-> PersonalName
domainpart <-> X.400 domain
(e.g., spaces are not allowed in an RFC822 domain, GQ and CN cannot
be directly mapped into localpart, DDAs of another type than RFC-
822), such attributes, together with the PN, are left-hand-side
encoded. The domainpart must still be mapped according to the mapping
rule as far as possible. This probably needs some examples:
C=nl; ADMD=ade; PRMD=tlec; O=owe; OU=you; S=plork; GQ=jr
->
/S=plork/GQ=jr/@you.owe.tlec.nl
C=nl; ADMD=ade; PRMD=tlec; O=owe; OU=spc ctr; OU=u; S=plork
->
"/S=plork/OU=u/OU=spc ctr/"@owe.tlec.nl
Note that in the second example, 'O=owe' is still mapped to a
subdomain following the natural hierarchy. The problems start with
the space in 'OU=spc ctr'.
3.3.2.2.2. RFC822 -> X.400
As an example, assume the following mapping rule is defined:
tlec.nl#PRMD$tlec.ADMD$ade.C$nl#
Then the address '[email protected]' :
localpart
sdom3
sdom2
sdom1
top-level-domain
[email protected]
would be mapped as follows.
The localpart 'plork' is mapped to 'S=plork', see chapter 3.3.2.1.
The domain 'tlec.nl' is mapped according to the mapping rule:
S OU OU O PRMD ADMD Country
plork tlec ade nl
The remaining domains (owe.you) are mapped one by one following the
natural hierarchy: sdom2 is mapped to O, sdom3 is mapped to OU:
S OU OU O PRMD ADMD Country
plork tlec ade nl
owe you
Thus the mapped address is (in a readable notation):
C=nl; ADMD=ade; PRMD=tlec; O=you; OU=owe; S=plork
Had there been any left-hand-side encoded SAs in the localpart that
didn't represent a complete mnemonic O/R address, the localpart would
be mapped to those SAs. E.g.,
"/S=plork/GQ=jr/OU=u/OU=spc ctr/"@owe.tlec.nl
->
C=nl; ADMD=ade; PRMD=tlec; O=owe; OU=space ctr;
OU=u; S=plork; GQ=jr
This is necessary to reverse the special use of localpart to left-
hand-side encode certain attributes. See 3.3.2.2.1.
You might ask yourself by now why such rules are needed at all. Why
don't we just use map1 in the other direction? The problem is that a
symmetric mapping function (a bijection) would indeed be ideal, but
it's not feasible. Asymmetric mappings exist for a number of reasons:
- To make sure that uucp addresses etc. get routed over local
gateways.
- Preferring certain address forms, while still not forbidding
others to use another form. Examples of such reasons are:
- Phasing out old address forms.
- If an RFC822 address is mapped to ADMD= ; it means that
the X.400 mail can be routed over any ADMD in that
country. One single ADMD may of course send out an
address containing: ADMD=ade; . It must also be possible
to map such an address back.
So we do need mapping rules from RFC822 to X.400 too. The table
containing the listing of all such mapping rules, which is
distributed to all gateways world-wide, is normally referred to as on
which software your are using) are:
'rfc2or'
'mapping 2'
'map2'
'table 2'
'R2X'
If the RFC822 localpart and/or domainpart contain characters that
would not immediately fit in the value of a PN attribute (! % _), the
mapping algorithm falls back to DDA mapping. In this case, the SAs
that will be used are still determined by mapping the domainpart
according to the mapping rule. In our case:
100%[email protected]
->
DD.RFC-822=100(p)user(a)work.tlec.nl;
C=nl; ADMD=ade; PRMD=tlec; O=work
If no map2 rule can be found, a third table of rules is scanned: the
gateway table. This table has the same syntax as mapping table 2, but
its semantics are different. First of all, a domain that only has an
entry in the gateway table is always mapped into an RFC822 DDA. For
a domain that is purely RFC822 based, but whose mail may be relayed
over an X.400 network, the gateway table associates with such a
domain the SAs of the gateway to which the X.400 message should be
routed. That gateway will then be responsible for gatewaying the
message back into the RFC822 world. E.g., if we have the gateway
table entry:
gov#PRMD$gateway.ADMD$Internet.C$us#
(and we assume that no overruling map2 rule for the top level domain
'gov' exists), this would force all gateways to perform the following
mapping:
[email protected]
->
DD.RFC-822=bush(a)dole.gov;
C=us; ADMD=Internet; PRMD=gateway
This is very similar to the default DDA mapping, except the SAs are
those of a gateway that has declared to be responsible for a certain
RFC822 domain, not those of the local gateway. And thus, this
mechanism helps avoid the third party problem discussed in chapter
3.2.2.
The table containing the listing of all such gateway rules, which is
distributed to all gateways world-wide, is normally referred to as
the 'gateway table'. Other commonly used filenames (also depending on
which software your are using) are:
'rfc1148gate' {From the predecessor of RFC1327, RFC1148}
'gate table'
'GW'
Only when no rule at all (map2 or gateway rule) is defined for a
domain, the algorithm falls back to the default DDA mapping as
described in 3.3.1.2.
3.4. Table co-ordination
As already stated, the use of mapping tables will only function
smoothly if all gateways in the world use the same tables. On the
global level, the collection and distribution of RFC1327 address
mapping tables is co-ordinated by the MHS Co-ordination Service:
SWITCH Head Office
MHS Co-ordination Service
Limmatquai 138
CH-8001 Zurich, Europe
Tel. +41 1 268 1550
Fax. +41 1 268 1568
RFC822: [email protected]
X.400: C=ch;ADMD=arcom;PRMD=switch;O=switch;S=project-team;
The procedures for collection and distribution of mapping rules can
be found on the MHS Co-ordination Server, in the directory
"/procedures". Appendix D describes how this server can be accessed.
If you want to define mapping rules for your own local domain, you
can find the right contact person in your country or network (the
gateway manager) on the same server, in the directory "/mhs-
services".
3.5. Local additions
Since certain networks want to define rules that should only be used
within their networks, such rules should not be distributed world-
wide. Consider two networks that both want to reach the old top-
level-domain 'arpa' over their local gateway. They would both like to
use a mapping 2 rule for this purpose:
TLec in NL: arpa#PRMD$gateway.ADMD$tlec.C$nl#
SWITCH in CH: arpa#PRMD$gateway.ADMD$switch.C$ch#
(You may have noticed correctly that they should have defined such
rules in the gateway table, but for the sake of the example, we
assume they defined it in mapping table 2. This was the way things
were done in the days of RFC987, and many networks are still doing
it this way these days.)
Since a mapping table cannot contain two mapping rules with the same
domain on the left hand side, such 'local mappings' are not
distributed globally. There exists a RARE draft proposal [13] which
defines a mechanism for allowing and automatically dealing with
conflicting mapping rules, but this mechanism has not been
implemented as to date. After having received the global mapping
tables from the MHS Co-ordination Service, many networks add 'local'
rules to map2 and the gateway table before installing them on their
gateways. Note that the reverse mapping 2 rules for such local
mappings _are_ globally unique, and can thus be distributed world-
wide. This is even necessary, because addresses that were mapped with
a local mapping rule may leak out to other networks (here comes the
third party problem again...). Such other networks should at least be
given the possibility to map the addresses back. So the global
mapping table 1 would in this case contain the two rules:
PRMD$gateway.ADMD$tlec.C$nl#arpa#
PRMD$gateway.ADMD$switch.C$ch#arpa#
Note that if such rules would have been defined as local gate table
entries instead of map2 entries, there would have been no need to
distribute the reverse mappings world-wide (the reverse mapping of a
DDA encoded RFC822 address is simply done by stripping the SAs, see
3.3.1.1.).
3.6. Product specific formats
Not all software uses the RFC1327 format of the mapping tables
internally. Almost all formats allow comments on a line starting with
a # sign. Some examples of different formats:
RFC1327
# This is pure RFC1327 format
# table 1: X.400 -> RFC822
#
PRMD$tlec.ADMD$ade.C$nl#tlec.nl#
# etc.
# table 2: RFC822 -> X.400
#
arcom.ch#ADMD$arcom.C$ch#
# etc.
EAN
# This is EAN format
# It uses the readable format for X.400 domains and TABs
# to make a 'readable mapping table format'.
# table 1: X.400 -> RFC822
#
P=tlec; A=ade; C=nl; # tlec.nl
# etc.
# table 2: RFC822 -> X.400
#
arcom.ch # A=arcom; C=ch;
# etc.
PP
# This is PP format
# table 1: X.400 -> RFC822
#
PRMD$tlec.ADMD$ade.C$nl:tlec.nl
# etc.
# table 2: RFC822 -> X.400
#
arcom.ch:ADMD$arcom.C$ch
# etc.
Most R&D networks have tools to automatically generate these formats
from the original RFC1327 tables;, some even distribute the tables
within their networks in several formats. If you need mapping tables
in a specific format, please contact your national or R&D network's
gateway manager. See chapter 3.4.
3.7. Guidelines for mapping rule definition
Beware that defining mapping rules without knowing what you are doing
can be disastrous not only for your network, but also for others. You
should be rather safe if you follow at least these rules:
- First of all, read this tutorial;.
- Avoid local mappings; prefer gate table entries. (See chapter
3.5)
- Make sure any domain you map to can also be mapped back;.
- Aim for symmetry.
- Don't define a gateway table entry if the same domain already
has a map2 entry. Such a rule would be redundant.
- Map to "ADMD=0;" if you will not be connected to any ADMD for
the time being.
- Only map to "ADMD= ;" if you are indeed reachable through
_any_ ADMD in your country.
- Mind the difference between "PRMD=;" and "[email protected];" and make
sure which one you need. (Try to avoid empty or unused
attributes in the O/R address hierarchy from the beginning!)
- Don't define mappings for domains over which you have no
naming authority.
- Before defining a mapping rule, make sure you have the
permission from the naming authority of the domain you want
to map to. Normally, this should be the same organisation as
the mapping authority of the domain in the left hand side of
the mapping rule. This principle is called 'administrative
equivalence'.
- Avoid redundant mappings. E.g., if all domains under 'tlec.nl'
are in your control, don't define:
first.tlec.nl#O$first.PRMD$tlec.ADMD$ade.C$nl#
last.tlec.nl#O$last.PRMD$tlec.ADMD$ade.C$nl#
always.tlec.nl#O$always.PRMD$tlec.ADMD$ade.C$nl#
but rather have only one mapping rule:
tlec.nl#PRMD$tlec.ADMD$ade.C$nl#
- Before introducing a new mapped version of a domain, make
sure the world can route to that mapped domain;.
E.g., If you are operating a PRMD: C=zz; ADMD=ade; PRMD=ergo;
and you want to define the mapping rules:
map1: PRMD$ergo.ADMD$ade.C$zz#ergo.zz#
map2: ergo.zz#PRMD$ergo.ADMD$ade.C$zz#
Make sure that ergo.zz (or at least all of its subdomains) is
DNS routeable (register an MX or A record) and will be routed
to a gateway that agreed to route the messages from the
Internet to you over X.400.
In the other direction, if you are operating the Internet
domain cs.woodstock.edu, and you want to define a mapping for
that domain:
map2: cs.woodstock.edu#O$cs.PRMD$woodstock.ADMD$ .C$us#
map1: O$cs.PRMD$woodstock.ADMD$ .C$us#cs.woodstock.edu#
Make sure that C=us; ADMD= ; PRMD=woodstock; O=cs; (or at
least all of its subdomains) is routeable in the X.400 world,
and will be routed to a gateway that agreed to route the
messages from X.400 to your RFC822 domain over SMTP. Within
the GO-MHS community, this would be done by registering a
line in a so-called domain document, which will state to
which mail relay this domain should be routed.
Co-ordinate any such actions with your national or MHS'
gateway manager. See chapter 3.4.
4. Conclusion
Mail gatewaying remains a complicated subject. If after reading this
tutorial, you feel you understand the basics, try solving some real-
life problems. This is indeed a very rewarding area to work in: even
after having worked with it for many years, you can make amazing
discoveries every other week........
Appendix A. References
[1] Postel, J., "Simple Mail Transfer Protocol", STD 10, RFC821,
USC/Information Sciences Institute, August 1982.
[2] Crocker, D., "Standard for the Format of ARPA Internet Text
Messages", STD 11, RFC822, University of Delaware, August 1982.
[3] Mockapetris, P., "Domain Names - Concepts and Facilities", and
"Domain Names - Implementation and Specification", STD 13, RFCs
1034 and 1035, USC/Information Sciences Institute, November
1987.
[4] Kille, S., "Mapping Between X.400 and RFC822", RFC987, UK
Academic Community Report (MG.19), UCL, June 1986.
[5] Braden, R., Editor, "Requirements for Internet Hosts --
Application and Support", STD 3, RFC1123, USC/Information
Sciences Institute, October 1989.
[6] Postel, J., Editor, "Internet Official Protocol Standards", STD
1, RFC1500, USC/Information Sciences Institute, August 1993.
[7] Chapin, L., Chair, "The Internet Standards Process", RFC1310,
Internet Activities Board, March 1992.
[8] Kille, S., "Mapping between X.400(1988) / ISO 10021 and RFC
822", RFC1327 / RARE RTR 2, University College London, May
1992.
[9] Kille, S., "X.400 1988 to 1984 downgrading", RFC1328 / RARE RTR
3, University College London, May 1992.
[10] Plattner, B., and H. Lubich, "Electronic Mail Systems and
Protocols Overview and Case Study", Proceedings of the IFIP WG
6.5 International working conference on message handling systems
and distributed applications; Costa Mesa 1988; North-Holland,
1989.
[11] Houttuin, J., "@route:100%[email protected], a practical guide to MHS
configuration", Top-Level EC, 1993, (not yet published).
[12] Alvestrand, H., "Frequently asked questions on X.400", regularly
posted on USEnet in newsgroup comp.protocols.iso.x400.
[13] Houttuin, J., Hansen, K., and S. Aumont, "RFC1327 Address
Mapping Authorities", RARE WG-MSG Working Draft, Work in
Progress, May 1993.
[14] "COSINE MHS Pocket User Guide", COSINE MHS Project Team 1992.
Also available in several languages from the MHS Co-ordination
Server:/user-guides. See Appendix D.
[15] Grimm, R., and S. Haug, "A Minimum Profile for RFC987", GMD,
November 1987; RARE MHS Project Team; July 1990. Also available
from the MHS Co-ordination Server:/procedures/min-rfc987-
profile. See Appendix D.
[16] CCITT Recommendations X.400 - X.430. Data Communication
Networks: Message Handling Systems. CCITT Red Book, Vol. VIII -
Fasc. VIII.7, Malaga-Torremolinos 1984.
[17] CCITT Recommendations X.400 - X.420. Data Communication
Networks: Message Handling Systems. CCITT Blue Book, Vol. VIII
- Fasc. VIII.7, Melbourne 1988.
Appendix B. Index
<<Only available in the Postscript version>>
Appendix C. Abbreviations
ADMD Administration Management Domain
ARPA Advanced Research Projects Agency
ASCII American Standard Code for Information Exchange
ASN.1 Abstract Syntax Notation One
BCD Binary-Coded Decimal
BITNET Because It's Time NETwork
CCITT Comite Consultatif International de Telegraphique et
Telephonique
COSINE Co-operation for OSI networking in Europe
DFN Deutsches Forschungsnetz
DL Distribution List
DNS Domain Name System
DoD Department of Defense
EBCDIC Extended BCD Interchange Code
IAB Internet Architecture Board
IEC International Electrotechnical Commission
IESG Internet Engineering Steering Group
IETF Internet Engineering Task Force
IP Internet Protocol
IPM Inter-Personal Message
IPMS Inter-Personal Messaging Service
IPN Inter-Personal Notification
ISO International Organisation for Standardisation
ISOC Internet Society
ISODE ISO Development Environment
JNT Joint Network Team (UK)
JTC Joint Technical Committee (ISO/IEC)
MHS Message Handling System
MOTIS Message-Oriented Text Interchange Systems
MTA Message Transfer Agent
MTL Message Transfer Layer
MTS Message Transfer System
MX Mail eXchanger
OSI Open Systems Interconnection
OU(s) Organizational Unit(s)
PP Mail gatewaying software (not an abbreviation)
PRMD Private Management Domain
RARE Reseaux Associes pour la Recherche Europeenne
RFCRequest for comments
RTC RARE Technical Committee
RTR RARE Technical Report
SMTP simple mail transfer protocol
STD Internet Standard
TCP Transmission Control Protocol
UUCP Unix to Unix CoPy
Appendix D. How to access the MHS Co-ordination Server
Here is an at-a-glance sheet on the access possibilities of the MHS
Co-ordination server:
address:
RFC822: [email protected]
X.400: S=mhs-server; OU1=nic; O=switch; P=switch; A=arcom;
C=CH
body
help # you receive this document
index ['directory'] # you receive a directory listing
send 'directory''filename' # you receive the specified file
FTP
address: Internet: nic.switch.ch
account: cosine
password: 'your email address'
Interactive
address: Internet: nic.switch.ch
address: PSPDN: +22847971014540
address: EMPB/IXI: 20432840100540
account: info
directory: e-mail/COSINE-MHS/
FTAM
address: Internet: nic.switch.ch
address: PSPDN : +22847971014540
address: EMPB/IXI: 20432840100540
address: ISO CLNS: NSAP=39756f11112222223333aa0004000ae100,
TSEL=0103Hex
account: ANON
gopher
address: Internet: nic.switch.ch
Security Considerations
Security issues are not discussed in this memo.
Author's Address
Jeroen Houttuin
RARE Secretariat
Singel 466-468
NL-1017 AW Amsterdam
Europe
Tel. +31 20 6391131
Fax. +31 20 6393289
RFC822: [email protected]
X.400: C=nl;ADMD=400net;PRMD=surf;O=rare;S=houttuin