Summary
This paper assumes a very basic understanding of AMHS and ATN Directory.
For those unfamiliar with AMHS and ATN directory, a simple introduction
is given in the Isode introduction
to the Aviation industry. A simple explanation of the nature of
the ATN Directory and its deployment in support of AMHS is given in
the Isode White Paper Deploying the
ATN Directory with AMHS: What you can do now.
This paper looks at how an AMHS end application, such as an AMHS Terminal
sending and receiving flight plans, will utilize and benefit from the
directory.
Model of Directory Usage
An AMHS user is someone using an end application communicating over
AMHS. A typical example is an AMHS
Terminal used by an operator to handle flight planning. This AMHS
user will communicate with other AMHS users. The basic model of AMHS
use of ATN Directory is that AMHS users are represented by entries in
the ATN Directory. For example:

This example shows a simple directory hierarchy (Directory
Information Tree (DIT)). Each blue circle represents an entry in the
directory, which may contain information about the object in question.
The top level of the DIT is countries represented by standard two letter
codes (e.g., C=FR is France). Under the fictional C=TY, the CAA is TYCAA,
and this is represented in the DIT as O=TYCAA (O= means “organization”).
Users are listed under O=TYCAA, with Common Name (CN=) from names. This
example illustrates users named descriptively and with an eight letter
AFTN name which can be represented by the following directory names:
- CN=YYYABCDE, O=TYCAA, C=TY
- CN=Operations Manager, O=TYCAA, C=TY
This hierarchical naming enables every AMHS user in the
world to have a unique directory name.
Each CAA will operate directory servers containing information
on their own users. These servers will be interconnected to form the
global ATN Directory, so that users of the ATN Directory will be able
to see entries for all users (and not just those for the local CAA).
When sending an AMHS message to that other user, the AMHS
client application (the AMHS Terminal in this example) is aware of the
recipient's entry in the ATN Directory. The AMHS client will use the
directory in two ways:
- To help locate the user (recipient discovery).
- To validate the recipient and recipient capabilities.
This second function is the one that is mandated by AMHS.
Both functions are described in more detail below.
Verification Layer
Recipient verification is the key AMHS use of the ATN
Directory. This function will not be usually be directly visible to
the end user. It will take place prior to a message being submitted
to the X.400 Message Transfer Service. There are two elements of this
verification
Recipient Name Verification
This core check is to determine the validity of the recipient,
by finding the recipient's entry in the ATN Directory. The benefit of
making this check is that it will detect messaging address errors prior
to a message being sent. This will mean that an operator can immediately
correct the address. If this check was not done, the error would show
up later when a non-delivery report is sent back. The name verification
will improve overall service resilience, by performing this check prior
to submission.
Recipient Capability Verification
The AMHS client will also read information from the entry,
about the capabilities of the recipient AMHS User. Some examples:
- Maximum message size. This directory attribute is defined in X.400.
It allows the client to check that the message being sent is not too
large to be handed by the recipient.
- Extended ATS Support. This directory attribute is defined in the
AMHS SARPs (Standards and Recommended Practices) and indicates that
the recipient supports the Extended ATS Service of AMHS, and so it
is safe to use features of this service. If this attribute is not
present, messages should be restricted to the Basic ATS Message Service.
- MTCU Capabilities. This directory attribute is defined in the AMHS
SARPSs and indicates that the recipient is not a direct user of AMHS,
but messages are sent though an MTCU (Message Transfer and Conversion
Unit) to an AFTN or CIDIN recipient. Information in this attribute
describes the capabilities of the MTCU, which will constrain the details
of messages that may safely be sent to this recipient.
- BUFR Support. This directory attribute should be defined as part
of a specification for carrying BUFR (Meteorological) messages over
AMHS. It would enable the sender to ensure that BUFR messages were
only sent to BUFR capable recipients.
The key point about recipient capability checking is that
it will enable the sender to be confident that messages can be correctly
handled by the intended recipient. It will enable new services based
on AMHS (that will initially only be supported by some recipients) to
be introduced in a robust manner.
When the Recipient is Not in the Directory
It may be that the message recipient does not have a directory
entry that is accessible by the AMHS client. Reasons for this may be:
- The recipient is in a country (CAA) where no directory is operated.
- The recipient’s CAA operates a directory, but it is not connected
to the directory being accessed by the AMHS Client. This would be
an unfortunate situation, and illustrates the importance of CAAs interconnecting
directories when they interconnect AMHS links.
In this situation, recipient verification by the AMHS
client is not possible, and so the message must be sent without verification.
In this situation, AMHS requires use of the Basic ATS Message Service
(no advanced features). This is desirable, to maximize interoperability.
It can be seen that directory deployment is a necessary pre-requisite
to deployment of advanced ATS applications.
Recipient Discovery
Recipient discovery takes place prior to recipient verification.
This use of directory is visible to the end user, and is a clear benefit
of using the ATN Directory. An AMHS client may obtain recipient information
in three different ways, discussed in detail below:
- Local (No Directory)
- Global (ATN) Directory
- Remote Recipients in the Local Directory
Local (No Directory)
There is no requirement for an AMHS client to get recipient
information from a directory. This information may be remembered, or
obtained from a local online address book (e.g., by the user scrolling
to and selecting a recipient).
Global (ATN) Directory
The ATN Directory is an ideal mechanism for the AMHS client
to select a recipient. The AMHS client can simply browse or search the
ATN directory, using the natural Country/Organization(CAA)/User hierarchy.
When the ATN Directory is used for recipient discovery,
verification is performed at the same time.
A good client works in a way that makes it easy for the
user to select frequently used recipients. It may also cache information
locally, to prevent repeated access to the directory to obtain the same
information.
Remote Recipients in the Local Directory
When a CAA deploys a directory, a first step will be to
provide the CAA's own part of the ATN Directory. This data will be managed
by the CAA in conjunction with AMHS configuration and will add immediate
value to the local AMHS deployment.
As the local ATN Directory is connected to other directories,
this will provide local users with access to more and more recipients
directly through the ATN Directory.
It will be desirable for local users to have convenient
access to information about recipients not (yet) in the global ATN directory.
This could be achieved by local means (as described above). An alternate
approach is to use the directory as "address book" storage
for remote recipients not in the ATN directory. This will have the benefits
of a client/server approach, and in particular will allow the address
book to be shared between multiple local users.
Conclusions
This paper has shown that the ATN Directory adds value
to an AMHS Deployment. Because of the usefulness of having information
on local users available in the directory, we believe that when a CAA
deploys AMHS, it should deploy its local part of the ATN Directory at
the same time. When CAAs interconnect AMHS, they should also interconnect
their directories, using a combination of directory chaining (X.500
Directory System Protocol) and directory replication (X.500 Directory
Information Shadowing Protocol) to optimize overall performance.