[ih] The UCLA 360/91 on the ARPAnet/Internet

Noel Chiappa jnc at mercury.lcs.mit.edu
Sun May 13 09:39:06 PDT 2012


    > From: Vint Cerf <vint at google.com>

    > The network working group did nco, telnet, FTP, eventually smtp,
    > although the latter might have been post-1982 and for Internet and not
    > Arpanet (?).

I thought SMTP came out of MSGROUP, or perhaps some hybrid of MSGGROUP and
the old NWG? (I vaguely recall chit-chat from other hackers in the MIT AI Lab
about it.) It _definitely_ was not from the TCP/Ip group(s).

None of RFCs 772, 780, 788 or 821 says anything definitive about where it
came from, alas. There are some (much earlier) RFCs about nail which I didn't
delve into which may help on this.

(Amusingly, 772 expands 'NCP' to "Network Control Protocol", which is what I
for many years thought/assumed it was - but it turns out that in the
beginning, it means 'Network Control Program'. Nice to know there is some
formal backup for 'Protocol'... :-)

SMTP was always intended for use with both NCP and TCP, and saw most of its
early use under NCP.


    > From: dave.walden.family at gmail.com

    > Regarding a second ARPANET host at UCLA, and other places, e.g. MIT: my
    > memory is that a lot of ARPA NET traffic was intra IMP site.

Indeed - especially between the ITS machines at MIT, which had a shared file
system across the ARPANET (and built into the OS, so all apps automatically
had access to it), so that any place you could name/use a file on the machine
you were on, you could just as easily name/use one on one of the others. So
that tended to generate a lot of traffic.

I recall seeing some numbers for packets at MIT, which split the data up into
inter-site and intra-site, but alas I don't think I have access to them any
more (there were weekly reports from BBN, IIRC).

	Noel



More information about the Internet-history mailing list