Aucbcory.220

Aucbcory.220
net.followup
utcsrgv!utzoo!decvax!ucbvax!ARPAVAX:CAD:ESVAX:Cory:cc-treas
Mon Mar 8 00:45:28 1982
Terminal Security Problem
I am a member of the Berkeley Division Academic Senate Committee
on Computing. We make Computing policy for the Berkeley Campus of the
University of California. The thrust of the problem has been described
accurately in Human-nets (as opposed to the news media) as involving
block send mode commands on intelligent terminals. I personally heard
about it last spring. The Committee was informed in October, and we got
the introduction to a paper by Donn Parker (and some other notable whose
name escapes me at the moment) at the same time.
InfoWorld, a micro newspaper, which, no doubt, some of you read,
got wind of it and put it in their January issue. It has gone wild in the
media since then. The L.A. Times article, which someone typed in and sent
over USENET (Unix UUCP news to the ARPAnauts), is the worst. Let me
emphasize that none of our systems have been breached in the manner
described. The Committee has discussed a number of measures that we
can take to remove the threat, and the best one is simply to not buy the
terminals with said feature. It is the opinion of the Committee that there
is no good reason for the block send command to exist as a sequence that
the Host computer can send. (If someone does have a good explaination,
I'd love to hear it!) It would be rather hard to modify the opsys to
exclude the lead-in characters from terminal-to-terminal communication
for ALL terminals. Hazeltines use ~, adm3a's use ESC. I've seen wierder
ones too. Essentially you'd have to outlaw the whole ASCII (or EBCDIC)
character set to fix things. And it still would be possible for someone
to mail you a letter with such things in it. So it is really up to the
terminal manufacturers to remedy the problem.
One thing that no one has mentioned yet is terminals with programmable
keys (i.e. keys that may be programmed by the user to send an arbitrary
sequence of characters). These terminals pose the same threat. I can
program such a terminal with all kinds of bombs, and wait for some
unsuspecting user to press one of them. Or, if the terminal is so equipped,
I can send it a code remotely to tell it to retransmit it's programmable keys.
I think the televideo 950 is one such.
Sorry for the long letter, but I had to get this off my chest. If
anyone has any questions, please mail to me directly.

Erik E. Fair
Representative to the Academic Senate Committee on Computing
Computer Science Undergraduate Association Treasurer
Cory.cc-treas@Berkeley (ARPA)
ucbvax!ucbcory!cc-treas (UUCP)

-----------------------------------------------------------------
gopher://quux.org/ conversion by John Goerzen <[email protected]>
of http://communication.ucsd.edu/A-News/

This Usenet Oldnews Archive
article may be copied and distributed freely, provided:

1. There is no money collected for the text(s) of the articles.

2. The following notice remains appended to each copy:

The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996
Bruce Jones, Henry Spencer, David Wiseman.