Introduction
This document contains my personal and highly partisan opinions on both
X.400 and the Internet mail protocols. As someone who has had to work
extensively with the OSI protocol family, this is not an impartial, objective
evaluation, but merely my own personal opinions on the relative strengths
and weaknesses of the two protocol families.
This, and the fact that the standards are drawn by committee is, in my opinion
the greatest weakness of OSI. Too many people are involved in the
standardisation process, many of whom will have a greater interest in defining
a standard that they can make money out of, rather than defining a good
standard.
This may seem sensible to those working in the PTT's and the communications
industry, but when you define a bad standard, it is quite likely that
people (not being stupid) will not use it.
In my opinion, there has been too much emphasis on defining standards for
services that can be charged for, rather than defining standards for services
of sufficient quality and economy that people would wish to pay for them.
The main feature of X.400 that is fairly immediatly apparent from the standard
is it's generality and extensibility. The standard was designed to last through
changes in the capabilities of technology - as happened with the advent of
personal computers with support for multiple media.
The way that other protocols (such as X.435, mentioned eariler), can be
added to the X.400 family of standards is a significant advantage which
may allow X.400 to benefit from new communications applications which
may arise in the future.
One of the advantages that X.400 has at the moment is that including
support for multimedia messages was one of the original design criteria.
Thus, the handling of multimedia mail messages is handled gracefully
in X.400, since it is an integral part of the standard, rather than something
that was added later.
As stated above, I think that one of the main problems with the OSI standards
is that too much influence has been exterted by PTT's to define a standard
that will increase their profits.
More concrete and less subjective complaints would be such things as the
standards documents themselves - in terms of readability, organisation,
notation and language, they are exceedingly poorly written.
Some people might claim that they are meant to be a formal definition of the
standards and are not meant to satisfy the above criteria, but this argument
breaks down when non-conformant or incompatible software is written because
developers could not understand, or failed to understand correctly, obtuse
or poorly expressed definitions.
In addition, other standards bodies (and not just 'informal' ones such as
the Internet Engineering Task Force (IETF)) can produce documents which,
if not exactly aesthetically pleasing, are nonetheless significantly more
readable and helpful to the developer. (The American National Standards
Institue for example).
Cubersome and over-complicated solutions - again, this is a matter of ones point
of view. From the point of view of the Internet "camp", the OSI protocol family
is too inefficient and cumbersome to be successful in the real world. From
the point of view of the OSI camp, they are merely paying the price for
safety and reliability. My own opinion lies somewhere in between - some
of the extra overhead in X.400 (and OSI) is justifiable given what it
tries to accomplish - some of it is not. A prime example of unjustifiable
overhead is the BER (Basic Encoding Rules) for ASN.1. Even at the time
when the standard was defined, it must have been obvious for any who
cared to look that the BER system for encoding objects was hideously
inefficient. More effiecient encoding schemes can reduce the size
of the encoded objects to a fraction of the size of those produced by
BER. This is significant and unneccessary overhead.
Not enough emphasis on implementability and simplicity. Again we see
a difference in philosophy between the two camps - according to the
Internet camp, a standard is useless unless it can be implemented
straight away. The OSI camp tends to aim for "ideal" systems, which
may not be suited for current technology. (In terms of speed of system
necessary, resources consumed etc.).
In addition to this, there is a large amount of stable and well tested
software (easy to accomplish with a user base of millions), much of it free.
There is an emphasis throughout on effieciency and low cost; making Internet
mail inexpensive both in terms of physical and financial resources. This has
lead to a huge proportion of the worlds' academic institutions joining the
Internet, and hence using Internet mail as their primary electronic mail
system.
One of the main problems with the Internet mail protocols is that they were
not designed to be extensible. This lack of generality and extensibility
causes problems when the time comes to add new features and extensions
to a standard not designed with these in mind.
This is particularly apparrent with the definition of the MIME extensions,
driven by the need for multi-media support for Internet mail. The MIME
standard, while not a bad one, is very much the result of having to graft
additional functionality onto a protocol not really suitable for it. While
the result has been quite successful (MIME, allthough by no means used by
the majority of Internet mail users, is probably more widely used than
X.400 itself), it lacks the elegance and simplicity of it's predecessors.
Another very important issue which the internet mail protocols do not
really address satisfactorily is that of security. Security is a major
concern for many "heavyweight" communications users such as Banks and
Financial Institutions.
Again, solutions do exist (such as PGP (Pretty Good Privacy) Public Key
Encryption), but it is again a question of grafting additional pieces
of functionality onto an existing standard - it is not integral to the
standards themselves, and has certainly not been a primary design
consideration at even more basic levels than encryption.
What becomes immediately apparent from any attempt to compare the two standards
is that they are trying to solve different problems. The IETF define standards
to solve problems that exist at the present time, and which can be implemented
immediately (and usually are).
The CCITT are concerned with defining standards which will last for a
considerable period of time and which will direct a considerable amount of
expenditure by the PTT's and others in the Telecommunications industry on
both hardware and software.
The result is that the IETF come up with simple solutions which are typically
implemented within a very short period of time, and software for which is
often distributed without charge accross the Internet.
The CCITT typically comes up with heavyweight solutions which take several years
to implement, which are very expensive, but for which support is available
from the supplier.
For the majority of users, who will only ever want to send text mail messages,
with the occasional encapsulated image or piece of audio, X.400 is an
inefficient, expensive and overcomplicated solution. However, for those
for highly sophisticated users for whom a high level of security, reliabilty
and integral support for multimedia messages (such as the military, banks
and certain companies), X.400 may well be the best solution available.
X.400
X.400 is a standard defined by the CCITT and ISO. While these are supposedly
international standards bodies, they are very much directed by special interest
groups such as the PTT's of the member countries, and to a lesser extent by
major players in the computing industry who have a (commercial) interest in
communications.
X.400 - Advantages
X.400 - Disadvantages
Internet Mail
Internet Mail - Advantages
As previously stated, one of the main advantages of the Internet mail
protocols is the massive installed user base. This has overcome many
problems - including the lack of provision for interfacing with other
protocols. The huge number of people on the Internet has lead to
a large incentive for both users and suppliers of other systems
to build and install gateways between their own systems and
the Internet.
Internet Mail - Disdvantages
Not designed to work with other standards. This has to a large extent been
overcome by the sheer numbers of users on the Internet.
Comparison
Links