NETWORK CHECKOUT
A July 1971 scoreboard of which early ARPANET sites had passed which software checks.
Harslem, E. and J.F. Heafner, "NETWORK CHECKOUT", RFC 193, DOI 10.17487/RFC0193, July 1971, <https://www.rfc-editor.org/info/rfc193>.
@misc{rfc193,
series = {Request for Comments},
number = 193,
howpublished = {RFC 193},
publisher = {RFC Editor},
doi = {10.17487/RFC0193},
url = {https://www.rfc-editor.org/info/rfc193},
author = {E. Harslem and J.F. Heafner},
title = {{NETWORK CHECKOUT}},
pagetotal = 2,
year = 1971,
month = jul,
} RFC text
Network Working Group 14 July 1971
Request for Comments: 193 E. Harslem - Rand
NIC 7138 J. Heafner - Rand
NETWORK CHECKOUT
As of 15 July we had contacted or been contacted by the sites
listed below. The matrix indicates the elements of Network software
we have verified at each site.
Telnet-
NCP- ICP User Server like Log-on Accessed
NCP-1 107 80 Telnet Telnet Protocol Procedure Services
-----------------------------------------------------------------
SDC | x | | | | | | | |
-----------------------------------------------------------------
UCLA-91 | | x | | x | x | | | |
-----------------------------------------------------------------
UCLA-S7 | x | x | x | x | x | | x | x |
-----------------------------------------------------------------
SRI-10 | | x | | | | x | x | |
-----------------------------------------------------------------
UCSB | x | x | x | | | | x | x |
-----------------------------------------------------------------
MIT-10 | | x | | x | | | | |
-----------------------------------------------------------------
MIT-645 | |par-| | | | | | |
| |tial| | | | | | |
-----------------------------------------------------------------
LL-67 | x | | x | | x | | | |
-----------------------------------------------------------------
BBN-10 | | x | | | | x | x | x |
-----------------------------------------------------------------
UTAH | x | x | x | | | x | | |
-----------------------------------------------------------------
Notes
SRI-10: The Telnet-like protocol differs from Telnet in
that lines are terminated by X'OD' rather than
X'ODOA'. John Melvin will have it fixed by
the time you read this.
RFC 193
LL-67: Joel Winett telles me NCP, ICP, EBCDIC, Telnet and
access to a virtual 360 are working. Due to hard-
ware difficulties we have not verified this.
BBN-10: Same difference as SRI with respect to Telnet.
Under certain conditions, the ICP generates a
zero byte size. Ray Tomlinson tells me these
idiosyncracies will be taken care of in about
a week. Three sites have commented that TENEX
assigns the same local socket to multiple connec-
tions, when acting as a server. Ray says that
this is the design -- not an error. BBN TENEX
is currently being used productively via the
Net, for some non-Network related problems.
UTAH: Our tests with UTAH were pre-TENEX. TENEX is
currently being installed at UTAH.
MIT-645: Our understanding is that most software is
working; we have not verified this.
[ This RFC was put into machine readable form for entry ]
[ into the online RFC archives by Paul Williams 6/97 ]
AI ELI5
On this page
§1What Was Checked
As of 15 July 1971, the Rand authors had contacted, or been contacted by, ten sites: SDC, UCLA-91, UCLA-S7, SRI-10, UCSB, MIT-10, MIT-645, LL-67, BBN-10 and UTAH. The matrix marks which elements of network software they had verified at each site.
The columns are NCP-1, two ICP columns labelled 107 and 80, user Telnet, server Telnet, a Telnet-like protocol, the log-on procedure and accessed services. An x means the element was verified. MIT-645 is marked only as partial.
The authors made a grid with the computers down the side and the tests across the top, and put an x where a computer passed.
§2Results by Site
UCLA-S7 has the most marks of any site, seven of the eight columns, followed by UCSB with five, including the log-on procedure and accessed services. BBN-10 has the Telnet-like protocol, log-on and accessed services marked. SDC has only NCP-1.
For LL-67, the authors were told by Joel Winett that NCP, ICP, EBCDIC, Telnet and access to a virtual 360 were working, but hardware difficulties meant they had not verified this. For MIT-645, they understood most software worked but had not verified it. UTAH had been tested before TENEX, which was then being installed.
Some computers passed nearly every test, some passed one or two, and for a couple the authors had only been told it worked and had not seen it themselves.
§3Problems Noted
At SRI-10 the Telnet-like protocol ended lines with X'0D' rather than X'0D0A', which differs from Telnet. John Melvin was to have it fixed. BBN-10 showed the same difference, and under certain conditions its ICP generated a zero byte size. Ray Tomlinson said these quirks would be fixed in about a week.
Three sites commented that TENEX assigns the same local socket to multiple connections when acting as a server. Ray said this is the design, not an error. BBN TENEX was already being used productively via the Net for some non-network problems.
A couple of computers ended their lines in a slightly different way than the rule book says, and the owners promised to fix it soon.