[TUHS] Early Internet work

Noel Chiappa jnc at mercury.lcs.mit.edu
Sat Feb 18 02:51:43 AEST 2017

    > OK, we're starting to get through all the clearances needed to release
    > the non-MIT Unix systems

We have now completed (as best we can) the OK's for the 'BBN TCP/IP V6 Unix',
and I finally bestirred myself to add in the documentation I found for it,
and crank out a tarball, available here:


It includes all the documentation files I found for the Rand and BBN code (in
the ./doc directory); included are the original NROFF source to the two Rand
publications about ports, and several BBN reports.

This is an early TCP/IP Unix system written at BBN. It was not the first
TCP/IP Unix; that was one done at BBN in MACRO-11, based on a TCP done in
MACRO-11 by Jim Mathis at SRI for the TIU (Terminal nterface Unit).

This networking code is divided into three main groups. First there is
code for the kernel, which includes IPC enhancements to Unix, including
Rand ports, as well as further extensions to that done at BBN for the
earlier TCP - the capac() and await() calls. It also includes a IMP
interface driver (the code only interfaced to the ARPANET at this point in
time). Next, TCP is implemented as a daemon which ran as a single process
which handled all the connections. Finally, other programs implement
applications; TELNET is the only one provided at this point in time.

The original port code was written by Steven Zucker at Rand; the extensions
done at BBN were by Jack Haverty. The TCP was mostly written by Mike
Wingfield, apparently with some assistance by Jon Dreyer. Dan Franklin
apparently wrote the TELNET.

Next, I'll be working on the MIT-CSR machine. That's going to take quite a
while - it's a whole system, with a lot of applications. It does include FTP,
SMTP, etc, though, so it will be a good system for anyone who wants to run V6
with TCP on a /23. We'll have to write device drivers for whatever networking
cards are out there, though.


More information about the TUHS mailing list