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

[Robert Braden]

The side discussion of the two IBM 360/91s at UCLA would
seem to have little to do with Internet history. However, I
want to point out that the "other" 360/91 at UCLA was in
fact a relatively early (1971) ARPAnet host, and it became
an Internet host after the Big Switch in Jan 1983. It supported
Telnet, FTP (a rich version, suited to the baroque IBM file
system), and SMTP.  This machine was distinguished by being
one of two 91s with 4M bytes of main memory (the other
belonged to a secret facility in the DC area). Its TCP/IP
implementation was one of the 5 experimental
implementations of TCP[/IP] put together by members of
the Internet* working group . I wrote the code in the 1977-
1983 time frame. The initial code V 2 of TCP. When
IP was split off  (V 2.5, as I recall), I applied a virtual scalpel
to the TCP code to split off the IP layer.

  I left UCLA for ISI in 1986 before I got around to adding Van's
  congestion control to TCP, sadly.

One other historic note: during 1981-1983 I maintained
  the /91 TCP/IP code remotely  over the ARPAnet from UCL
  in London. Kirstein had a TIP on the ARPAnet,
Remote debugging over the ARPAnet was certainly not new,
but I could probably can claim a distance record at the time ;-)

Bob Braden

*Perhaps called the Catenet working group at the time?