rfc333.txt

著名的RFC文档,其中有一些文档是已经翻译成中文的的.

Network Working Group                                       Bob Bressler
Request for Comments: 333                           MIT/Dynamic Modeling
NIC # 9926                                                    Dan Murphy
Category: C9 (experimentation)                                 BBN/TENEX
Obsoletes: 62                                                Dave Walden
Updates: none                                                    BBN/IMP
                                                             15 May 1972


        A PROPOSED EXPERIMENT WITH A MESSAGE SWITCHING PROTOCOL


CONTENTS

   Introduction ..................................................  1
   Some Background ...............................................  2
   References ....................................................  3
   MSP Specification .............................................  4
   Issue .........................................................  8
   Message Header ................................................ 10
   Examples ...................................................... 15
   TELNET ........................................................ 16
   The Information Operator ...................................... 16
   Unique Port Numbers ........................................... 20
   Flow Chart .................................................... 23
   MSP Variations ................................................ 25
   Appendix ...................................................... 26

INTRODUCTION

   A message switching protocol (MSP) is a system whose function is to
   switch messages among its ports.

   For example, there is an implementation of an MSP in each Interface
   Message Processor.  We believe that the effective utilization of
   communications networks by computer operating systems will require a
   better understanding of MSPs.  In particular, we feel that Network
   Control Programs (NCPs), as they have been implemented on the ARPA
   Computer Network (ARPANET), do not adequately emphasize the
   communications aspects of networking -- i.e., they reflect a certain
   reluctance on the part of systems people to move away from what we
   term "the stream orientation".  We propose, as an aside the network
   development using the current NCPs, to rethink the design of NCP-
   level software beginning with a consideration of MSPs.

   The thrust of this note is to sketch how one would organize the
   lowest level host-host protocol in the ARPANET around MSPs and how
   this organization would affect the implementation of host software.



Bressler, et al.            Experimentation                     [Page 1]

RFC 333          MESSAGE SWITCHING PROTOCOL EXPERIMENT          May 1972


SOME BACKGROUND

   Over the past several weeks there has been considerable informal
   discussion about the possibility of implementing, on an experimental
   basis, in several of the ARPA Network Host Computers, NCPs which
   follow a protocol based on the concept of message switching rather
   than the concept of line switching (see the parenthetical sentence in
   the first paragraph of page 6 of NIC document 8246, Host/Host
   Protocol for the ARPA Network).  Party to this discussion have been
   Bob Bressler (MIT/Dynamic Modeling) Steve Crocker (ARPA), Will
   Crowther (BBN/IMP), Tom Knight (MIT/AI), Alex McKenzie (BBN/IMP), Bob
   Metcalfe (MIT/Dynamic Modeling), Dan Murphy (BBN/TENEX), Jon Postel
   (UCLA/NMC), and Dave Walden (BBN/IMP).

   Several interesting points and conclusions have been made during this
   discussion:

      1. Bressler has implemented a message switched interprocess
         communication system for the Dynamic Modeling PDP-10 and has
         extended it so it could be used for interprocess communication
         between processes in the Dynamic Modeling PDP-10 and the AI
         PDP-10.  He reports that it is something like an order of
         magnitude smaller than his NCP.

      2. Murphy has noted that a Host/Host protocol based on message
         switching could be implemented experimentally and run in
         parallel with the real Host/Host protocol using some of the
         links set aside for experimentation.  Further, Murphy has noted
         that if this experimental message switching protocol were
         implemented in TENEX, a number of (TENEX) sites could easily
         participate in the experiment.

      3. It is the consensus of the discussants that Bressler should
         take a crack at specifying a message switching protocol* and
         that if this specification looked relatively easy to implement,
         a serious attempt should be made by Murphy and Bressler to find
         the resources to implement the experimental protocol on the two
         BBN TENEX and the MIT Dynamic Modeling and AI machines.

      4. MSP was chosen as the acronym for Message Switching Protocol,
         and links 192-195 were reserved for use in an MSP experiment.



   -------------
   *This note fulfills any obligation Bressler may have incurred to
   produce an MSP specification.




Bressler, et al.            Experimentation                     [Page 2]

RFC 333          MESSAGE SWITCHING PROTOCOL EXPERIMENT          May 1972


   We solicit comments and suggestions from the Network Working Group
   with regard to this experiment.  However, although we will very much
   appreciate comments and suggestions, because this is a limited
   experiment and not an attempt to specify a protocol to supersede the
   present Host/Host protocol for the ARPA Network, we may arbitrarily
   reject suggestions.

REFERENCES

   Familiarly with the following references will be helpful to the
   reading of the rest of this note.

      1) NIC document 8246, HOST/HOST PROTOCOL FOR THE ARPA NETWORK

      2) NIC document 9348 on the Telnet Protocol

      3) NIC document 7101, OFFICIAL INITIAL CONNECTION PROTOCOL,
         DOCUMENT # 2

      4) a system of interprocess communication in a resource sharing
         computer network, CACM, April, 1972.

   Reference 4 is a revision of RFC 62.  We strongly suggest the reader
   be familiar with reference 4 before he attempts to read the present
   RFC; a reprint of reference 4 is attached as an appendix.


























Bressler, et al.            Experimentation                     [Page 3]

RFC 333          MESSAGE SWITCHING PROTOCOL EXPERIMENT          May 1972


MSP SPECIFICATION

   Our MSP is essentially a generalization of the interprocess
   communication system outlined in Section 3 of the fourth reference.
   (Henceforth, if we are required to mention the interprocess
   communication system presented in Section 3 of reference 4, we shall
   call it "the IPC".)  For two processes to communicate using the MSP,
   the process desiring to send must in some sense execute a SEND and
   the process desiring to receive must in some sense execute a RECEIVE.
   The SEND and RECEIVE, in effect, rendezvous somewhere and
   transmission is allowed to take place.  With the RECEIVE are
   specified (among other things) a FROM-TO-PORT-ID, a TO-PORT-ID, and a
   RENDEZVOUS HOST.  With SEND are specified a from-port-id, a to-port-
   id, a rendezvous Host, and (possibly) some data to be transmitted.
   Using SEND and RECEIVE, sending a message from a SENDER PROCESS to a
   RECEIVER PROCESS takes place as follows.  The sender process executes
   a SEND which causes an OUT-MESSAGE plus the specified data to be
   transmitted to the Host specified as the rendezvous Host in the SEND.
   Concurrently (although not necessarily simultaneously)the receiver
   process executes a RECEIVE which causes an IN-MESSAGE to be sent to
   the Host specified as the rendezvous Host in the RECEIVE.  At the
   rendezvous Host, OUT-messages and IN-messages are entered in a table
   called the RENDEZVOUS TABLE.  When an OUT-message and an IN-message
   are detected with matching to-port-id, from-port-id, and rendezvous
   Host, three things are done:  1)  the OUT-message plus the data is
   forwarded to the Host which was the source of the IN-message, 2)  the
   IN-message is forwarded to the Host which was the source of the OUT-
   message, and 3)  the IN-message and OUT-message plus the data are
   deleted from the rendezvous table in the rendezvous Host.

   The process is greatly simplified if the rendezvous Host is also
   either the send Host or receive Host.  Specific algorithms
   enumerating these sequences appear later in this note.

   To clarify the basic concepts, let us look at a case involving three
   Hosts, to which we shall give the names SND, RCV, and RNDZ.  At Host
   SND, process S is doing a send, and at Host RCV, process R is doing a
   receive.  Both specify rendezvous at Host RNDZ.













Bressler, et al.            Experimentation                     [Page 4]

RFC 333          MESSAGE SWITCHING PROTOCOL EXPERIMENT          May 1972


+--------------------+     +----------+     +--------------------+
|HOST SND            |     |          |     |HOST RCV            |
|                    |     |          |     |                    |
|                    |     |          |     |                    |
|       (PROCESS)    |     +----------+     |                    |
|       (   S   )    |         HOST         |                    |
|              \     |         RNDZ         |          (PROCESS) |
|              [DATA]|                      |          (  R    ) |
+--------------------+                      +--------------------+


Process S now executes a SEND with

     from-port-id = S, to-port-id = R, and rendezvous-Host = RNDZ.

Host SND then creates a table entry in its rendezvous table.

+-----------------------------------+
|HOST SND            MSP   _ _ _    |
|           ------------->|_ _ _|   |
|         /        ^      |_ _ _| <-|-------RENDEZVOUS
|        /         |      |_ _ _|   |         TABLE
|(PROCESS)         |                |
|(   S   )         +-- SEND (from=S to=R; rend=RNDZ)
|        \                          |
|         [DATA]                    |
+-----------------------------------+

Host SND now sends an "OUT" message with S's data to Host RNDZ.

  HOST SND                               HOST RNDZ
+------------+                    +---------------------------+
|         MSP|  "OUT" + DATA      |MSP  _____  RENDEZVOUS     |
|            |--------------------|--> |_ _ _| TABLE          |
|            |  from=S; to=R      | \  |_ _ _|                |
|            |                    |  \ |_ _ _|                |
+------------+                    |   \             __        |
                                  |    \---------->|  | DATA  |
                                  |                |__|BUFFER |
                                  |                           |
                                  +---------------------------+

   Concurrently process R at Host RCV executes a RECEIVE with from-
   port-id = S, to-port-id = R, and rendezvous-Host = RNDZ.  As above,
   Host RCV creates a table entry in its rendezvous table and sends an
   "IN" message to Host RNDZ (see following figure).





Bressler, et al.            Experimentation                     [Page 5]

RFC 333          MESSAGE SWITCHING PROTOCOL EXPERIMENT          May 1972


   (Don't panic now about buffering in an intermediate Host.  The time
   to panic is afer you've read and understood the rest of our
   arguments.)

     HOST RNDZ                          HOST RCV
+------------------------+       +-----------------------+