Front-end Protocol B6700 version
RFC 705

Document Type RFC - Unknown (November 1975; No errata)
Last updated 2013-03-02
Stream Legacy
Formats plain text pdf html bibtex
Stream Legacy state (None)
Consensus Boilerplate Unknown
RFC Editor Note (None)
IESG IESG state RFC 705 (Unknown)
Telechat date
Responsible AD (None)
Send notices to (None)
Network Working Group
Request for Comments:  705
NIC# 33644

                         FRONT - END PROTOCOL
                            B6700 VERSION

                           2 September 1975

This is a working document which has been developed as the specification 
and guideline for design of a Burroughs B6700 attachment to an ARPA-Style

The approach is to utilize a front-end processor with a new protocol for 
network operation.  That protocol, described herein, has been built upon
the concepts expressed by M.A. Padlipsky, et al, in NIC# 31117, RFC# 647.

This proposed, site-specific, FEP implementation is the work of Gerald 
Bailey and Keith McCloghrie of NSA and of David Grothe of ACC.  It has 
already sustained some corrections provided by MAP.  It will be helpful
if interested networkers will review and provide comments to us.

Comments to BRYAN@ISI.

Roland Bryan - ACC                                                      1

Network Working Group
Request for Comments:  705
Front-End Protocol: B6700 Version

                        ***WORKING DOCUMENT***

                          FRONT-END PROTOCOL


This document describes the protocol to be used for connecting a general-
purpose computer system (host) to an ARPANET-like network via a "front-end" 
computer.  The main body of the document is aimed at a reader who is not 
conversant with all the details of network protocols.  However, a paragraph
marked with [n], refers a reader familiar with network protocols to the
n-th item of Appendix A which will amplify that particular paragraph.  
Further information on the network protocols referred to in this document 
can be obtained from the Network Information Center.

Appendix B contains diagrams showing the transitions between the different
connection states.  Appendices C and D give the implementation details of 
this protocol in the Front-End and the Hosts.

This protocol is predicated upon the assumption that for each host, a line
protocol, at a lower level, will be established between the device-driver
modules in the Host and the Front-End, and that this line protocol provides
Front-End Protocol with error-free transmissions.

                             INTRODUCTION                               2

A host computer may be connected to a network for a variety of reasons.  
Network connection may be an attempt to expand the usefulness of the
Host to the community of users which it serves by making network resources
available to them.  Conversely, the services which the Host provides may 
be made available to a larger community of users, with the network providing 
the method of access to those services.

In order for members of a network community to communicate in an intelligent 
way, there must exist a set of protocols.  The implementation of these 
protocols in a host computer is typically called the Network Control Program
(NCP).  The size and complexity of the NCP is proportional to the number and 
complexity of protocols which it implements.  For an ARPANET like network,
both the number and complexity are substantial.

                       ***WORKING DOCUMENT***


RFC 705
Front-End Protocol

                        ***WORKING DOCUMENT***

A host which directly connects into the network must assume the responsibility
for implementing this set of protocols.  That is the "price of admission"
to become a network host.  It is not necessary to implement every protocol
and every option in every host, but even in the simplest case -- implementation
of an NCP is not a small task.  The intrusion into the normal operating
environment of the host is also not small.

An alternative method for network connection is to connect the host to some
intermediate processor, and in turn, directly connect that processor to the
network.  This approach is called "Front-Ending."  There are many arguments
which may be posed to justify a host connection to a network through a front-
end processor.  The most obvious being that the responsibility for 
implementation of the network protocols (the NCP) can be delegated to the
front-end (FE), thereby reducing the impact on the host.

The purpose of this document is not to justify Front-Ending as a philosophy,
but rather, to introduce a protocol for communications between a host and
a front-end processor which is providing it network access.  The Front End
Protocol (FEP) is intended to permit the host to make use of the network 
through existing protocols, without requiring that it be cognizant of the 
complexities and implementation detail inherent in their execution.

The FEP is sufficiently general to permit its implementation in the host 
to be in terms of the function the host is performing, or the services
which it is providing.  Of primary consideration in specification of FEP
was that it must provide the host with a sufficiently robust command 
repertoire to perform its network tasks, while buffering it from the
Show full document text