RFC 1204 EXPERIMENTAL EMAIL · AUTHENTICATION

Message Posting Protocol (MPP)

Protocol for personal computers to submit outgoing mail to a mail service host.

On rfc-editor.org ↗
PUBLISHED
February 1991
AUTHORS
S. Yeh, D. Lee
STREAM
Legacy
TRANSPORT PROTOCOL
TCP
ASSIGNED PORT
218
MAIN COMMANDS
USER, PASS, DATA, NOOP, QUIT
DOI
10.17487/RFC1204

RFC text

Contents
Original text · RFC Editor View on rfc-editor.org ↗
Network Working Group                                             S. Yeh
Request for Comments: 1204                                        D. Lee
                                              Netix Communications, Inc.
                                                           February 1991


                     Message Posting Protocol (MPP)

Status of this Memo

   This memo describes a protocol for posting messages from workstations
   (e.g., PCs) to a mail service host.  This RFC specifies an
   Experimental Protocol for the Internet community.  Discussion and
   suggestions for improvement are requested.  Please refer to the
   current edition of the "IAB Official Protocol Standards" for the
   standardization state and status of this protocol.  Distribution of
   this memo is unlimited.

1. INTRODUCTION

   Operating systems for personal computers do not provide a mechanism
   for user authentication.  However, such a mechanism is crucial for
   electronic mail system since authenticating message sender's identity
   is important in preventing mail forgery.  Hence, adding personal
   computers to an electronic mail network requires an agent (message
   posting server) to authenticate sender's identity and then submit
   mail to the message delivery system (e.g., Sendmail, MMDF) on behalf
   of the sender at a PC.  The Netix Message Posting Protocol is
   developed to be the interface between the message posting server and
   the PC (client).  The protocol is designed to use TCP and is based on
   the command and reply structures defined for Simple Mail Transfer
   Protocol (RFC 821) and File Transfer Protocol (RFC 959).

2. SPECIFICATIONS

2.1.  Command List

      USER <SP> <username> <CRLF>
      PASS <SP> <password> <CRLF>
      DATA <CRLF>
      NOOP <CRLF>
      QUIT <CRLF>

2.2.  Reply List

      220 Message Posting Service Ready.
      221 Closing Connection.
      250 Command OK.

      354 Enter mail, end with <CRLF>.<CRLF>
      451 Local error encountered.
      500 Command unrecognized.
      501 Argument syntax error.
      503 Illegal command sequence.
      530 Authentication Failure.
      550 Error.

2.3.  Command and Reply Descriptions

      USER <SP> <username> <CRLF>

         The USER command informs the message posting server about the
         username of the user trying to submit mail to the network.  The
         required argument for the USER command is a string specifying
         the message sender's username.

         The USER command can only be used under three conditions:

         - when the session with the message posting server has just
           started;

         - right after a message text (terminated by the "<CRLF>.<CRLF>"
           sequence) has been successfully submitted to the message
           posting server;

         - right after a USER command that gets the reply code 501.

         List of possible reply codes for the USER command:

         - 250   The username of the message sender has been accepted.

         - 451   Internal error has occurred in the message posting
                 server.

         - 501   Syntax error detected in the username argument.

         - 503   The USER command has been used under an inappropriate
                 condition (i.e., one that is not specified above).

         It is recommended that the message posting server should return
         250 even if the username is not recognized by the message
         posting server, as long as the username is syntactically
         correct.  This is an attempt to prevent the message posting
         server from releasing too much information about the user
         database.  Client should not be able to test the existence of a
         certain username.

      PASS <SP> <password> <CRLF>

         The PASS command is used to inform the message posting server
         about the password associated with the username previously
         specified.  The required argument for the PASS command is a
         string specifying the message sender's password.

         The PASS command can only be used under two conditions:

         - right after a USER command that gets the reply code 250;

         - right after a PASS command that gets the reply code 501.

         List of possible reply codes for the PASS command:

         - 250   The password has been accepted and verified to be
                 correctly associated with the username previously
                 specified.

         - 451   Internal error has occurred in the message posting
                 server.

         - 501   Syntax error detected in the password argument.

         - 503   The PASS command has been used under an inappropriate
                 condition (i.e., one that is not specified above).

         - 530   The password provided is not the one associated with the
                 username previously specified.

      DATA <CRLF>

         The DATA command is used to inform the message posting server
         to get ready to accept a mail message text.  No argument is
         expected.  (This command has the same meaning as the DATA
         command defined in RFC 821.)

         The DATA command can only be used under two conditions:

         - right after a PASS command that gets the reply code 250;

         - right after a mail message text has been successfully
           accepted from the client.

         List of possible reply codes for the DATA command:

         - 354   The message posting server is ready to accept the mail
                 message text.

         - 451   Internal error has occurred in the message posting
                 server.

         - 503   The DATA command has been used under an inappropriate
                 condition (i.e., one that is not specified above).

         Upon receiving the reply code 354 for the DATA command, the
         client should submit the mail message text to message posting
         server and terminate the text by the sequence "<CRLF>.<CRLF>"
         as defined in RFC 821.  If the message text includes the
         "<CRLF>.<CRLF>" sequence, then the sequence is replaced by the
         "<CRLF>..<CRLF>" sequence as defined in RFC 821.  The extra "."
         token will not be included in the final copy of the submitted
         message.

         Upon receiving the mail message text terminated by the
         "<CRLF>.<CRLF>" sequence, list of possible reply codes is:

         - 250   The mail message text has been successfully queued for
                 delivery.

         - 451   Internal error has occurred in the message posting
                 server and the mail message text has not been queued.

      NOOP <CRLF>

         The NOOP command does not cause any action to be performed by
         the message posting server.  Instead, it tests the status of
         the message posting server.  No argument is expected.

         The NOOP command cannot be used under one condition:

         - right after a DATA command that gets the reply code 354
           (i.e., when the message posting server is expecting the client
           to submit the mail message text).

         List of possible reply codes for the NOOP command:

         - 250   The message posting server has not encountered any
                 internal error.

         - 451   Internal error has occurred in the message posting
                 server during the current session.

      QUIT <CRLF>

         The QUIT command is used to terminate the session with the
         message posting server.  No argument is expected.

         The QUIT command can be used under any condition.  The message
         posting server should always return the reply code 221 for the
         QUIT command.

3. IMPLEMENTATION OF THE MESSAGE POSTING SERVER

   There are several issues to be considered when implementing the
   message posting server:

   - secured environment
   - port number assignment
   - handling of idle client
   - local/remote password database
   - message queuing
   - handling of message delivery failure

3.1  Secured Environment

   The message posting server is responsible for authenticating message
   senders and submitting mail to the message delivery system.  Hence,
   it should be running in a secured environment, such as running on a
   system (UNIX, VMS, MS-DOS) with well restricted physical and network
   access.

3.2  Port Number Assignment

   Port 218 is assigned for the Netix Message Posting Protocol.

3.3  Handling of Idle Client

   The message posting server should terminate a session if the client
   has been idle for too long, to release the resource allocated for the
   session.

3.4  Local/Remote Password Database

   To take advantage of existing password databases, such as the passwd
   file in UNIX, the message posting server can use FTP and POP3 to
   perform the username and password checking with the appropriate
   server.

   For network that does not have any password database, the message
   posting server should let the system administrator specify a local
   password file on the host that the message posting server is running.

3.5  Message Queuing

   The message posting server should attempt to submit accepted messages
   to the message delivery system as soon as possible.

3.6  Handling of Message Delivery Failure

   Failure in delivering messages should be handled by the message
   delivery system and the message posting server should not interfere.

4. REFERENCES

   [1] Postel, J., "Simple Mail Transfer Protocol", RFC 821,
       USC/Information Sciences Institute, August 1982.

   [2] Postel, J., and J. Reynolds, "File Transfer Protocol", RFC 959,
       USC/Information Sciences Institute, October 1985.

Security Considerations

   Security issues are discussed in section 3.1.

Authors' Addresses

   Shannon Yeh
   Netix Communications, Inc.
   15375 Barranca Parkway, Suite A-215
   Irvine, CA 92718

   Phone: (714) 727-9335

   Email: yeh@netix.com


   David Lee
   Netix Communications, Inc.
   15375 Barranca Parkway, Suite A-215
   Irvine, CA 92718

   Phone: (714) 727-9335

   EMail: dlee@netix.com

AI ELI5

On this page
PROBLEMPersonal computers lack operating-system-level user authentication mechanisms, which is essential for email systems to prevent mail forgery and ensure senders' identities are verified.
MECHANISMMPP defines a simple command-reply protocol over TCP that allows a PC to authenticate with a mail server using username and password, then submit messages for delivery to the wider email system.
INTEGRATION SOLUTIONThe protocol enables PCs to integrate into email networks by handling authentication at the server level before submission to mail delivery systems like Sendmail or MMDF.

§1The Problem and Solution

Personal computers running operating systems like MS-DOS do not provide operating-system-level user authentication. This creates a problem for email: the email system must authenticate a sender's identity to prevent mail forgery. The solution is to use an intermediary agent called a message posting server that authenticates the sender and then submits mail to the message delivery system on behalf of the sender.

The Netix Message Posting Protocol defines the interface between the message posting server (running on a trusted host) and the PC client. The protocol uses TCP and is based on command-reply structures from RFC 821 (SMTP) and RFC 959 (FTP).

Your computer at home does not have a secure way to prove who you are. So you log in to a trusted mail server, and it sends your mail for you.

§2Commands and Reply Codes

The protocol defines five main commands. USER informs the server of the username attempting to send mail, restricted to three conditions: session start, after successful message submission, or after a syntax error on a previous USER command. PASS provides the password associated with the previously specified username.

DATA instructs the server to accept mail message text, followed by the client sending lines of message content terminated by a bare period on its own line (matching RFC 821 convention). NOOP tests the status of the server without performing any action. QUIT terminates the session.

Reply codes follow SMTP conventions: 220 (service ready), 250 (command OK), 354 (ready for message), 451 (local error), 500 (unrecognized command), 501 (syntax error), 503 (illegal command sequence), 530 (authentication failure), and 550 (error).

The protocol has simple commands: tell your name, give your password, say you want to send a message, send the message, and hang up.

§3Server Implementation Considerations

The RFC identifies several implementation issues for message posting servers. Security is paramount: the server must run in a secured environment with restricted physical and network access. Port 218 is assigned for the protocol.

Idle client handling should terminate sessions if no activity occurs for too long to free resources. For username and password verification, servers can leverage existing password databases (such as UNIX passwd files) using FTP and POP3, or maintain a local password file specified by the system administrator. The server should accept any syntactically correct username and return 250 to avoid leaking information about which usernames exist.

Accepted messages should be queued and submitted to the message delivery system immediately. Message delivery failures should be handled by the message delivery system itself, not by the message posting server.

The server needs to be careful about security, not keep clients connected forever, and pass messages to the real mail system quickly.

§4Security and Operational Status

The MPP protocol addresses one specific security problem: authenticating the originator of a message before it enters the mail system. By requiring username and password authentication on the message posting server, the server can verify that a PC user is who they claim to be before accepting their mail.

RFC 1204 remains an Experimental specification rather than a standard. While it provided an early approach to the problem of mail authentication from untrusted clients, later solutions like SMTP authentication extensions (SMTP AUTH) became the standardized approach for client authentication in the Internet email ecosystem.

The main security idea is to check who you are before letting you send mail. Later standards provided better ways to do this.

§5References