The RISKS Digest
Volume 4 Issue 91

Thursday, 28th May 1987

Forum on Risks to the Public in Computers and Related Systems

ACM Committee on Computers and Public Policy, Peter G. Neumann, moderator

Please try the URL privacy information feature enabled by clicking the flashlight icon above. This will reveal two icons after each link the body of the digest. The shield takes you to a breakdown of Terms of Service for the site - however only a small number of sites are covered at the moment. The flashlight take you to an analysis of the various trackers etc. that the linked site delivers. Please let the website maintainer know if you find this useful or not. As a RISKS reader, you will probably not be surprised by what is revealed…

Contents

Electromagnetic Interference in Japan
Lindsay F. Marshall
Risk of Inappropriate Technology to Prevent Password Overwrite
Paul Stachour
Passwords and Statistics
Earl Boebert
Why Cellular phones at the Indy 500?
Robert Adams
Information Security Products and Services Catalog by NSA
Kurt F. Sauer
Re: TRW "Credentials"
John R. Levine
Phalanx Schmalanx
PGN
Mike Trout
Torkil Hammer
Laser guides
Jon A. Tankersley
Re: Risks of running Risks
Jeff Woolsey
Will Martin
Re: Computer thefts
David Phillip Oster
Info on RISKS (comp.risks)

Electromagnetic Interference (in Japan)

"Lindsay F. Marshall" <lindsay%kelpie.newcastle.ac.uk@Cs.Ucl.AC.UK>
Thu, 28 May 87 9:40:09 BST
CHIPS ARE DOWN OVER ELECTRONIC POLLUTION  (The Guardian, 26 May 1987)

Japan is engulfed in an "electronic smog" which has caused deaths and injuries,
and jammed an airport radar system, according to recent findings.

Electronic smog occurs when electromagnetic waves from equipment like personal
computers and electronic game machines "escape" and trigger other machines.  An
electromagnetic wave can also be caused by a mere spark.  An electric spark
from a crane operating in a valve plant set off a lathe-operating robot in
1982 killing an assembly-line worker.

A simple household device like a television aerial booster can have dire
consequences.  Osaka International Airport's radar screens were jammed by
electromagnetic waves from a nearby television aerial booster.  The fault
and its cause were discovered.  But the Ministry of post and
Telecommunications (MPT), which has set up a group to investigate the
problem, admits that an air crash could have occurred.

Forty-two people were injured last September when two cars on a
roller-coaster crashed.  The MPT suspects electromagnetic waves from an
unknown source were to blame.  It says that communications at a busy railway
switching point have been jammed by waves from television game machines.  Train
doors have inadvertently opened several times.  The MPT wants manufacturers to 
redesign their electronic goods so that the waves cannot escape.

Lindsay F. Marshall, Computing Lab., U of Newcastle upon Tyne, Tyne & Wear, UK
JANET: lindsay@uk.ac.newcastle.cheviot  ARPA: lindsay%cheviot.newcastle@ucl-cs
PHONE: +44-91-2329233                   UUCP: <UK>!ukc!cheviot!lindsay

   [Remember the item in RISKS-4.89 attributing at least six Japanese 
   robot-related deaths to electromagnetic interference.   PGN]
       [[By the way, Lindsay's message said "lather-operating robot".
       As this case was not a close shave, I assume it was a typo.]]


Risk of Inappropriate Technology to Prevent Password Overwrite

<Stachour@HI-MULTICS.ARPA>
Thu, 28 May 87 07:52 CDT
  In Risks Digest 4.86, PGN comments on the over-long password, and the
article by Bill Young in IEEE Symposium on Security and Privacy, April 1987.
I agree that most of us do a bad job of mapping our specifications to our
implementations, and Bill does an excellect job of pointing out one way to
overcome this risk.
  However, I believe that the case Bill points out is but one case of a very
general problem that has been solved in a different manner, in a much less
risky way.  Specifically, the specification that was violated in the
implementation can be summarized as "A data area is overwritten by a method
that should have no access to that data area."  This over-writing problem is
quite general: it happens for arrays (such as when strings are implemented
as array of characters), for improperly based structures, and other places.
  The particular error cited by Bill Young could not have happened if the
implementation had been in a language such as PL/I or Ada, where
over-running the bounds of an array is a required run-time check (in the
cases where the compiler cannot determine at compile-time that the
assignment is not 100% safe) instead of in a language like C where all the
effort is on the programmer, and no help is given to her by the language.
Such checks are clearly not new technology, since Multics (written in PL/I)
has been doing such for over 20 years.  Nor is the technology new to
hardware, since the Burroughs B5500-series and MCP (written in Algol) has
also been checking for a similar period.
  One of the reasons for the reliability of Multics and MCP is that they do
NOT depend on perfect programmers, but use both the language run-time
(checking software descriptors) and hardware (segmentation, rings, ...)  to
check for common overwrite errors and other access violations.
  I would personally prefer to put my faith in a compiler generating
correct code and the run-time for the language checking descriptors properly 
than putting my faith in every coder of every array-reference in every
program, even where all such programs have been "proved" to be correct,
simply due to the sheer size and difficulty of such proofs, and the high
probabiliy of an error in the proof, or in the configuration-control
of different versions (one proved correct, a different one installed)
of the software, or some other error that makes the proof "non-valid".

  This leads to my question:
    What RISK do we bring on to ourselves (both personally and
professionally) when we ALLOW inapproprate technologies to be used
(or in many cases, forced by others) that thereforce create unreliable,
non-robust software, when both the hardware and software technology
to prevent many of the problems have existed long enough that every
reasonable person working in software should be aware of them.
[I know that many people today only study the things of the current
generation, done simply and incorrectly for simple systems,
and don't understand much else, but that's another question.]

Paul Stachour, Honeywell & University of Minnesota, Stachour@UMN-CS.EDU


Passwords and Statistics

<Boebert@HI-MULTICS.ARPA>
Tue, 26 May 87 10:50 CDT
From the Computer Shopper, May 1987:

Password Snatcher — RS-232 Data tap.  Lets you actually "see" the data
being sent or received on an RS-232 line.  Connect a terminal or microcomputer
to the tap connector to capture data, or connect a serial printer to get a 
hard copy.  Jumpers allow for routing TD, RD, or both to the tap.  $29.95.

         [Great jumpers.  A Computershopper must be like a grasshopper.  PGN]

Re the statistics on use of opposite-sex names for passwords:  The best
rebuttal to this kind of statistical argument came from the redoubtable
John W. Campbell:  The laws of population growth tell us that
approximately half the people who were ever born in the history of the
world are now dead.  There is therefore a 0.5 probability that this
message is being read by a corpse.


Why Cellular phones at the Indy 500?

Robert Adams <adams@littlei.UUCP>
Tue May 26 15:00:45 1987
    While watching the Indianapolis 500 on TV this Sunday, I saw them do a
feature on one of the car crews that were using a celluar phone to talk to
the driver on the track.  You see, most crews use some sort of CB or
shortwave set to talk between the pit and the driver and the TV announcers
are always talking about what they overheard on the radios.  This one car
had a celluar phone and the crew would phone the driver to discuss things.
This seemed really strange to me until I realized that the use of the phone
meant that no one could legally listen in on their conversations.
    Everyday someone discovers a new way to use that law.

Robert Adams, Intel Corp., ISO Systems Development, Hillsboro, OR


Information Security Products and Services Catalog by NSA

"Kurt F. Sauer" <ks%a.cs.okstate.edu@RELAY.CS.NET>
Thu, 28 May 87 3:09:56 CDT
RISKS Readers who have a professional or personal interest in information
security on a practical level might be interested to learn that I have just
received my copy of the April 1987 Information Security Products and Services 
Catalogue, prepared by the National Security Agency.  There are no protective 
markings on the document, and (since they're VERY CLEAR when things *aren't* 
public-domain) I presume it is available on request by writing to

    Director, National Security Agency, 9800 Savage Road, 
    Fort George G. Meade MD 20755-6000

Anyway, it's in interesting compendium of companies who provide
communications and information systems security devices and services.
According to some accompanying document(s), "...this catalogue will be
distributed quarterly to [organizations] who received [certain lists]
previously...  Plans are under way to request that they be provided through
the Government Printing Office on a subscription basis."

The document is in 5 parts and lists companies and equipments thus:

    (1)  Endorsed Cryptographic Products List
    (2)  NSA Endorsed Data Encryption Standard (DES) Products List
    (3)  Protected Services List
    (4)  Evaluated Products List
    (5)  Preferred Products List

Also included is a brief, but informative, description of categories of
information which requires security, information about purchase and restric-
tions on the purchase of these devices and services.

Happy hunting!   Kurt F. Sauer, Director of Operations, Decision Studies Group,
  Inc./OP, Post Office Box 701318, Tulsa, OK 74170-1318,  Tel +1 918 749 0893


Re: TRW "Credentials"

John R. Levine <johnl@ima.ISC.COM>
Mon, 25 May 87 12:57:07 EDT
Ron Heiby wrote in RISKS-4.89 about a report in Insight magazine on TRW's new
"Credentials" that gives you access to your TRW credit record for $35/year.

The current issue of Forbes also reports on this new so-called service.
Forbes points out that TRW is required by law to provide a copy of one's
credit record for free any time credit is denied because of a credit report,
and for a nominal reproduction fee at any other time.  They express
incredulity that people seem willing to pay $35 for what they could already
get for free.  In addition, TRW encourages the Credentials customers to add
extra information voluntarily to their credit records under the dubious
theory that this will help them to get credit in the future.  People seem to
think that because their info is in TRW's computer it will be more credible.*

John Levine, ima!johnl or Levine@YALE.somethingorother

* - Not a pun, no matter what you may think.


More on the Stark

Peter G. Neumann <Neumann@CSL.SRI.COM>
Tue 26 May 87 11:25:41-PDT
An article by Molly Moore of the Washington Post appeared in the SF
Chronicle, 26 May 1987, pp. 13 and 14.  It adds lots more details on the
Stark and its equipment.  The computers were programmed to give a LOUD alarm
that the missiles had been launched, even when NOT in the mode to fire the
Phalanx automatically.  Perhaps the radar scans only in the direction that
the Phalanx was pointing?  (We have already noted reports that the computers
were not even working at the time, although there is also a denial by the
Captain.)  On the other hand, the fact that the use of computers in this
context for automatic firing of the Phalanx is deemed unsafe is of great
interest to RISKS.  Reliance on continual preparedness in the defensive use
of unreliable systems with unprepared personnel seems quite risky.  

    [Thus the next item is relevant, for further background.  However,
    let me again remind contributors to cite your references carefully.
    I have rejected a few items that say "I vaguely remember seeing 
    somewhere that ... ."  Also, please try harder to avoid wild
    speculation.]


Phalanx Schmalanx

Mike Trout <rpics!brspyr1.BRS.Com!miket@seismo.CSS.GOV>
28 May 87 21:22:33 GMT
Regarding the recent USS Stark incident, much has been written about the
Phalanx anti-missile system mounted on the ship.   Generally, Pentagon
spokespeople--parrotted by the news media and net posters--have stated some
variation of the following:

"If only {choose one or more of the following}:

    A) the Phalanx had been switched to "automatic"
    B) the ship's stern, where the Phalanx is mounted, had been pointing
           toward the incoming missile(s)
    C) the crew had been alert/warned in time to properly use the Phalanx
    D) the Phalanx had been operating properly (crew allegations of
           assorted Phalanx malfunctions/down time/maintenance problems)

then the Phalanx would have easily blown the incoming missile(s) away and we
would all live happily ever after."

This universal faith in the Phalanx is a dangerous belief that I take exception
to.  Contrary to the way the Navy and the media talks, Phalanx is a very new
weapon that has NEVER been used in combat conditions.  

History tells us that military hardware testing conditions and combat "real
world" conditions probably don't even lie in the same universes.  US military
hardware testing, driven by profit motives and military career advancement, is
particularly atrocious.  Everything I've ever discovered about this situation
leads me to believe that to get a good idea of how a weapon will really
perform, the best strategy is to utterly IGNORE test results.

The things that the "military-industrial complex" (MIC) (why is that a taboo
term these days?--Eisenhower coined it) will do to test a weapon are truly
bizarre.  Drones (moving at the breakneck speed of 60 mph, without evasive
action) have been painted with gloss red enamel to absorb more laser energy,
to "prove" that lasers can shoot down planes.  Gigantic aluminum foil
bull's-eyes have been arranged on the ground so that the Pershing II's radar
could find its target.  Over 160 mobility tests of the M1 tank resulting in
the tank's breakdown were thrown out as "invalid", while the single test in
which the tank didn't break down was presented as the final result.  Every
single test of the GLCM cruise missile has been altered to cover up grossly
unacceptable navigational errors.

Most of the Phalanx testing data I'm aware of shows excellent results,
usually on the order of 80% of incoming targets hit.  But as I've stated
before, that may mean nothing at all.  One problem in the Phalanx testing
the MIC doesn't like to talk about is that even if the incoming missile is
hit, the hit takes place so close to the ship that the ship is still blasted
by the missile's wreckage.  The Phalanx's MAXIMUM range is only about 1.5
miles, and most anti-ship missiles cover that distance in just a few
seconds.  This problem becomes worrisome when you consider that of the four
Exocets that hit British ships off the Falklands, THREE were duds, including
the one that hit the HMS Sheffield.  But the devastation caused by a dud
missile was still severe enough to wreck or sink the ships hit.  Even
without an exploding warhead, an anti-ship missile is a large, heavy object
travelling at high speed carrying fuel tanks at least partially filled with
various formulas of highly volatile rocket fuel.  The force of the missile
mass times its velocity is transferred as heat to the ship, resulting in the
possibility of a devastating fire--the greatest danger to modern warships.
Even if an incoming missile is exploded by the Phalanx, most of the
missile's mass will still strike the ship.  The mass will, of course, be
dispersed by the explosion--the amount depending on how close to the ship
the missile was when the missile exploded--and the velocity will be reduced
as well.  Hopefully this will be enough to save the ship, but even so a ship
peppered by high speed flaming debris may be out of action for some time.

Also, remember that the Exocet is a fairly old design as anti-ship missiles go.
Most of the newer designs are faster, have better guidance, and might not be
duds.  The new Soviet SS-N-22 supposedly almost hides behind wavetops while
travelling at better than double the Exocet's speed.

The Phalanx's gun--a 20mm gatling--is an excellent, proven weapon that has
performed extremely well in real world plane-to-plane combat.  But
plane-to-plane combat is a different environment than ship-to-missile
combat.  I've never heard of a pilot shooting down a missile coming at his
plane with a 20mm gatling, or with any other weapon for that matter.  Years
ago, the US Army had a weapon called the "Chapparal", which was a 20mm
gatling mounted on an armored personnel carrier.  Admittedly, there weren't
many opportunities to try it out against enemy air, but it died a quiet
death.  I know some Army commanders complained about its short range and
high ammunition consumption.

This Phalanx faith also makes me worry about the AEGIS cruiser idea.  The
AEGIS is crammed with all kinds of anti-air and anti-missile stuff,
including multiple Phalanx systems, and nearly everybody thinks of it as
being able to instantly produce an inpenetrable shield.  But I can't help
thinking about an interview with the Second Fleet commander a couple of
years ago.  He said if a real war broke out, he'd send all his AEGIS
cruisers home.  He contented that its high-powered radar, tracking, ECM, and
weapons control electronics puts out such a blatant electromagnetic
signature that it would attract every Soviet plane, ship, and sub within
hundreds of miles.  Sort of like a bug light that attracts so many bugs that
despite its excess power, it can't kill the bugs fast enough and it clogs
and shorts out.

So let's not get carried away by the Phalanx.  It wasn't supposed to be a
perfect missile defense system, but the MIC has to justify all that expense.
It's probably useful as a last-ditch emergency measure, but thinking of it
as a missile umbrella leads to a "crutch mentality" that neglects more
useful missile defenses like air cover, anti-AIRCRAFT weapons, electronic
countermeasures/chaff, and evasive action.

Michael Trout (miket@brspyr1) =-=-=-=-=-=-= UUCP:ihnp4!dartvax!brspyr1!miket
BRS Information Technologies, 1200 Rt. 7, Latham, N.Y. 12110  (518) 783-1161


Re: Stark

Torkil Hammer <sdcsvax!sdcrdcf!psivax!torkil@ucbvax.Berkeley.EDU>
Tue, 26 May 87 13:33:17 PDT
American commentators have the curious notion that, when English-speaking
soldiers die, it must be explained as a result of human errors, but when
their foreign speaking counterparts go the ultimate way of soldiers, it is
considered a victory.  Compare the newsmedia coverage of the Sheffield and
Stark incidents with the Libya and Grenada ones.

It looks to me, that the only risk involved is one of messing with guys
packing Exocets, and ramifications depend on what language you speak.

The technical discussion in this newsgroup has barely mentioned that
Exocets are smart missiles with an unusually high rate of 'success', as
seen from the launching end.  Which translates to an equally high
rate of 'failure' as seen from the targets.

So a technical discussion on the failure of the ships' defenses must
logically be followed by a discussion on why the second missile arrived
without exploding.  Rather atypical for Exocets.  Usually they do.
Computer failure?

torkil hammer, Pacesetter Systems Inc., Sylmar, CA


Laser guides (RISKS-4.90)

"Jon A. Tankersley" <apctrc!zjat02@seismo.CSS.GOV>
27 May 87 14:57:43 GMT
Seems to me that the laser guided technology is about the same as the wire
guided technology of the Arab-Israeli War.  Wire guided TOW missles could
be easily defeated by spraying the general direction of origin.  Many
Israeli tanks ended the war with lots of wires draped over them.

The only 'smart' way that weapons can work is by adding intelligence, but
that can even cause problems (Berserker SF series by Saberhagen, etc.)

-tank-   Amoco Production Co, Tulsa Research Center    [Tanks.]


Re: Risks of running Risks ["One" the record]

<Jeff Woolsey <woolsey@nsc.NSC.COM<>
Tue, 26 May 87 09:47:16 PDT
  >    After 14 days (326 hours), your message has not yet been              |
  >fully delivered.  Attempts to deliver the message will continue           |
  >for 178956963 more days.  No further action is required by you.           V
  >   [********* = = = = = = = = = = = = = = = = = = = = = = = = = = = = = !!!!

That number is 0xAAAAAAA _+ 7 (1 week).   Even the machine shrieks in
astonishment at having to keep retrying that long.

Jeff Woolsey    National Semiconductor Corporation
...!nsc!woolsey -or- woolsey@nsc.COM -or- woolsey@umn-cs.ARPA


Re: Risks of running RISKS (RISKS 4.88)

Will Martin — AMXAL-RI <wmartin@ALMSA-1.ARPA>
Tue, 26 May 87 15:45:47 cdt
It was mail to me that generated that erroneous mailer barf message that
told you it was going to keep trying for 178 million-odd more days...

Sorry about that... Our net-host 750 has a flaky clock, even though it keeps
getting parts replaced by maintenance, and in that instance it crashed
briefly and came up with a system time thirteen days in the future.  Since,
at the same time, our main user machine had been down and mail to it was
sitting in the queue on the 750, the mail process on the 750 thought it was
two weeks since they had been queued and so merrily generated a great spew
of garbage mailer messages. I still don't know what is causing them to have
that enormous number of days displayed. We get all sorts of strange numbers
showing up in that field...

Anyway, when I noticed this, we cut off outgoing mail and I did a mass
delete of all message traffic with a "Waiting mail" subject line.
Unfortunately, it appears some slipped out before we caught them. Sigh...

If you see any again, please ignore them. (Though I'm sure you don't need
any more junk mail like that cluttering up your inbox.)
                                                          Will Martin


Re: Computer thefts (re: RISKS-4.82)

David Phillip Oster <oster%dewey.soe.Berkeley.EDU@Berkeley.EDU>
28 May 87 23:52:27 GMT
Apple computer sells a thing called a "security kit" that permanently
attaches a steel shackle to a Macintosh and Keyboard. With a steel
cable passing through both shackles, and padlocked to a water pipe, it
makes the Macintosh hard to steal. (Good for students.)

I don't know whether this feature was preserved in the new Macintosh
SE and Macintosh II, nor what it costs. 

The security kit does not protect the mouse, which can be easily
unscrewed and walked off with, but it should help cut down thefts of
the whole machine. 

Many third party vendors sell kits to lock computers down. Most work by
permanently fastening a baseplate to the desk under the machine with a tough
adhesive. Apple's version at least will not ruin a student's dorm's desk.
-- David
           [I hate to think of the flood resulting when 
           someone decides to cut the water pipe.  PGN]

Please report problems with the web pages to the maintainer

x
Top