References

Tanenbaum section 7.4 Electronic Mail

RFC 821 - Simple Mail Transfer Protocol

RFC 822 - Message format

RFC 2045 - MIME multipurpose Internet mail extensions - format of message bodies

RFC 2046 - MIME - media types

Electronic mail - recall why and what

Internet email - RFC 821 and 822

One of the earliest open application protocols - 1982

Very simple protocol for transferring human-readable mail messages.

Messages include envelope and content information.

OSI email - X.400

Open Systems alternative to Internet email.

A more complicated family of protocols over the OSI 7 layer stack, with

Little implemented, interoperability function largely replaced by Internet mail.

Email handling processes - agents

User agent interacts with user:

Transfer agent sends and receives messages over network - client-server, connection via port 25.

User Agents and Message Transfer

There are very many examples of user agents:

The user agent prepares the message from user interaction
including message transfer header fields;
expands user's address aliases

The message transfer agent (such as UNIX sendmail) acts as client, makes a connection with an agent acting as server at the destination host computer.

Messages are transferred with SMTP - simple mail transfer protocol
often via intermediate mail hosts
where address aliases may be expanded.

Messages are commonly stored at receiver for
later pickup by user agent

OR
transfer to further remote user using Post Office Protocol POP.

Format of email messages - original - RFC 822

Messages consist of

• header - a number of header lines
  Each has identifier : text

From: somebody

To: someone else

Subject: something that you want to write about

• blank line (CRLF)
• body - a number of lines

This format should be familiar to those who have looked at HTTP (see also lecture 5 in this series) - which is based on the RFC 822 standard.

Tanenbaum fig. 7-42, Tanenbaum fig. 7-43 Some RFC 822 headers

MIME - Multipurpose Internet Mail Extensions - RFC 2045

Added to basic message format to allow

• non-ASCII character sets - Arabic, Hebrew
• non-alphabetic character sets - multibyte representations: Japanese etc.
• non-ASCII character set extensions - non-English Roman alphabets etc with accents, diacritical marks etc
  and word processor character sets including special characters
• non-text data: images, audio, video
• arbitrary binary data for applications
• multiple parts of message in different formats

Content-type:

text/plain

image/gif

application/octet-stream

can also specify more interpretation information,
such as text/plain - character set in which the message is to be interpreted, using attribute=value notation added to type/subtype

text/plain charset=us-ascii

text/plain charset=iso8859-1

Headers added by MIME - Tanenbaum fig. 7-44

The MIME content types and subtypes - Tanenbaum fig. 7-45

Example

Tanenbaum fig 7-46 - multipart message

HTTP and MIME extensions to RFC822 message formats

These headers are also used in HTTP.

When appearing in the header of a HTTP response, the client (browser) is informed how to display the message body.

HTTP also adds further related headers, such as

Content-encoding:

[HTTP specification section 14.12, RFC 2068, HTTP/1.1]

specifies compression method used for body data

gzip

compress

Protocol for email message transfer RFC 821

The transfer protocol is simple.

Connection oriented - assumes error free transport service (TCP).

Client may transfer many messages in one session.

SMTP protocol sequence - Halsall "Data communication,s computer networks and open systems" 3rd edition, Addison Wesley 1992, fig 12-6

Example of message transfer - Tanenbaum fig. 7-47

Email transfer encoding

Needed to carry different types of data through the very low-ability RFC821 service of 7 bit ASCII characters only.

MIME header Content-transfer-encoding:

7bit - default if no header included

8bit

binary

quoted-printable - 7 bit characters plus hex-encoding for extension set

Now is winter=B4s ending=20

just around the corner.

base64 - groups of 24 bits grouped as 4x6 bits,
each 6 bit group represented by a single printable ASCII character
in A..Z, a..z, 0..9, +, /
where = and == mean a short group of 8 or 16 bits at the end.

AAUWBwACAAAAAAAAAAAAAAAAAAAAAAAAAAMAAAADAAAAPgAAABAAAAAJAAAATgAAACAA

AAACAAAAbgAAAR5BZ2VuZGEgMy85NyBKdWx5VzZCTk1TV0QBAP////8AAAAAAAAAAAAA

AAAAAAAAAAAAAAEAAAABAAAAAAAAAAAej0hu/+RwAS8ASG7/7Ehu//Bw+avtVE9Krv/s

Zwa4rv8QQWdlbmRhIDMvOTcgSnVseSACAAIAVzZCTk1TV0QAAAIAVzZCTk1TV0QAAAAA

AAAAAAAAAAAAAAAAAAAAAAAAAACv7/lAAABuAAAAAR5QIm7/4CNoAAQAGCBu//AgUCJu

/+AjUAAISgVnHkoGZxogeAK2IGgBGHAIwKgACGcKLy7/4E66BopYT1mPqNgmXy8KLwuo

3C8LcAE/AD8AqOEvCi8LLwqo5i8LqNlVjy8u/+AvLgAULwpwDavtNh9nDAxD/4BnBj1D

ABhgYgxD/4BmAAABAAAAAQAAAAAAAAAAHgKf7IwdBAAAABwAHv//