The Risks Digest

The RISKS Digest

Forum on Risks to the Public in Computers and Related Systems

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

Volume 22 Issue 64

Tuesday 18 March 2003

Contents

Apparently uncommanded rudder movement injures cruise passengers
Steve Peterson
Jeppesen GPS navigation database corruption
Mickey Coggins
California outage causes prescription mix-up
Richard Cook
Glitch let gamblers beat machines
M Taylor
Haywire ATM spits out extra cash
Fuzzy Gorilla
Beware the spelling checker
NewsScan
Recent worms punish bad passwords
David J. Aronson
Profile of a virus writer
NewsScan
Search engines making sensitive information easy to locate
Richard Moore
Benetton clothes to include tracking chip
Monty Solomon
CASPIAN calls for immediate worldwide boycott of Benetton
Monty Solomon
Federated network identity
Brian Seborg
Re: Computer crashes threaten hospital operations
Jonathan Kamens
Re: Monster electricity bill
Don Gingrich
Human protocol failure
Dawn Cohen
The Workshop on Rapid Malcode: WORM
Robert K. Cunningham
Info on RISKS (comp.risks)

Apparently uncommanded rudder movement injures cruise passengers

<Steve Peterson <steve@zpfe.com>>
Mon, 17 Mar 2003 11:52:03 -0600

I received an e-mail from relatives who are on a Holland America cruise
around the tip of South America.  Last Thursday night, while they were
eating dinner in the dining room, there was a sudden lurch and the ship went
into a hard right turn, listing over at an estimated angle of 20 degrees.
Every chair in the dining room toppled and people, dishes and food slid
everywhere.  At least two people were injured in the laundry when clothes
dryers topped on top of them.  The grand piano on the stage was flipped
over.

After about 5 minutes, the ship stopped turning.  Shortly after that, the
captain came on the PA system and announced that, when they turned off the
autopilot to enter the port at Buenos Aires, the rudder swung to the right
and could not be "unstuck."  Over two hours elapsed after the incident,
before the crew performed a count of the passengers.


Jeppesen GPS navigation database corruption

<Mickey Coggins <risks-spam-alert@int.ch>>
Fri, 14 Mar 2003 17:53:25 +0100

I just got this from the AOPA (Aircraft Owners and Pilots Association),
a truly outstanding organization, BTW.

  Jeppesen reports airspace boundary problems

  About 350 airspace boundaries contained in Jeppesen NavData are incorrect,
  the FAA has warned. The error occurred at Jeppesen after a software
  upgrade when information was pulled from a database containing 20,000
  airspace boundaries worldwide for the March NavData update, which takes
  effect March 20. Only a dozen are in the United States, including Chicago;
  Louisville, Kentucky; Fayetteville, North Carolina; Santa Ana, California;
  Las Vegas; Honolulu; Des Moines; and Oklahoma City. The error could cause
  pilot alerts to be given by GPS units too early or too late. Pilots are
  advised to use multiple sources of information, such as carrying paper
  charts (Jeppesen paper charts are unaffected by the problem), and
  contacting controlling agencies by radio to avoid airspace
  violations. Jeppesen has provided a searchable database of locations with
  airspace boundary errors on its Web site (
  http://www.aopa.org/epilot/redir.cfm?adid=1875 ). To search, click on the
  binocular icon and enter an airport identifier.  Jeppesen spokesman Mike
  Pound said the errors will be corrected "for the next possible release."
  The next release is scheduled for April 17.

The risks are pretty obvious.  Today if you fly your tiny aircraft into
restricted airspace in the USA, you can get shot down by agents of "The
Department of Homeland Security".

Many (most) pilots use this navigation data with their GPS as their primary
source of navigation information worldwide.

Please don't get me started on the absurdity of many of these airspace
restrictions in the first place...


California outage causes prescription mix-up

<"Richard Cook" <ri-cook@uchicago.edu>>
Tue, 18 Mar 2003 09:58:16 -0600

Thousands of patients could have received the wrong prescription drugs after
a power outage at Kaiser Permanente's computer center in Southern California
knocked the pharmacy's labeling system out of sync -- printing the wrong
labels on filled prescriptions.  There were no reports yet of patients
suffering from adverse reactions.  About 4,700 patients from Fresno to the
Oregon border were affected, including those ordering prescriptions by
telephone.  After the error was discovered on 14 Mar 2003, hospital
officials attempted to contact the affected patients, although by 17 Mar,
152 remained uncontacted -- including those for whom they had only PO-box
addresses.  [Source: Associated Press, 18 Mar 2003, PGN-ed; Also noted by
Danny Burstein]


Glitch let gamblers beat machines

<M Taylor <mctaylor@privacy.nb.ca>>
Thu, 13 Mar 2003 12:50:34 +0000

Some Nova Scotia video lottery terminal (VLT) players found a way to beat
machines last year, forcing the Atlantic Lottery Corp. to replace computer
chips in about one-third of their 3,538 VLTs in Nova Scotia and
Newfoundland.  About 20 locations appeared to have had the glitch exploited.
ALC doesn't know how much it lost, but noted the maximum payout for a single
spin is $500.  The problem was discovered in December 2002 after ALC
received calls from retailers who noticed frequent high payouts, and from
its own inspections.  Chips were replaced in January and February 2003.
[Glitch let gamblers beat machines, *The Chronicle-Herald*, Nova Scotia,
8 Mar 2003; PGN-ed]
  http://www.herald.ns.ca/

M Taylor  http://www.mctaylor.com/


Haywire ATM spits out extra cash

<"Fuzzy Gorilla" <fuzzygorilla@euroseek.com>>
Fri, 14 Mar 2003 18:54:43 -0500

Several bank customers in Fargo, North Dakota, had an automated teller
machine disperse more cash than they had requested.  (They turned it in.)
Apparently ``cold weather caused the ATM cash door to stick so some
customers who wanted to withdraw money could not get it.  The door opened
for other customers, who then wound up with their cash and the cash
belonging to the previous customers.''  [Source: AP, 14 Mar 2003; PGN-ed]
  http://story.news.yahoo.com/news
  ?tmpl=story2&u=/ap/20030314/ap_on_fe_st/haywire_atm


Beware the spelling checker

<"NewsScan" <newsscan@newsscan.com>>
Mon, 17 Mar 2003 09:42:19 -0700

A study at the University of Pittsburgh reveals that the ubiquitous spelling
checker software may be doing as much harm as good, when it comes to
writing.  In the study, 33 undergraduate students were asked to proofread a
one-page business letter -- half of them using Microsoft Word, with its
spelling- and grammar-checking features and the other half using only their
brains.  Without the software, students with higher SAT verbal scores made,
on average, five errors, compared with 12.3 errors made by students with
lower scores.  However, using the software, the two groups made about the
same number of errors -- 16 vs 17.  Dennis Galletta, a professor of
information systems at the Katz Business School, says people have come to
rely on spelling-checker software too completely. "It's not a software
problem, it's a behavior problem."  [AP 14 Mar 2003; NewsScan Daily, 17
March 2003; slight PGN-ed]
  http://apnews.excite.com/article/20030314/D7POQ7R80.html


Recent worms punish bad passwords

<"David J. Aronson" <dja2003@hotpop.com>>
Tue, 11 Mar 2003 10:20:36 -0500

  A spike in Internet traffic caused by a worm over the weekend can be
  largely blamed on bad passwords and poor security practices, said security
  experts on Monday. The Deloder worm, which spreads by communicating with
  Windows computers that have file sharing enabled, may have spread to
  perhaps as many as 10,000 systems using a list of 86 passwords to break
  into computers running Microsoft Windows NT, 2000 and XP.  [Source: MSNBC]
    http://www.msnbc.com/news/883415.asp:

The same article later says:

  The recent LovGate worm -- which appeared on the Internet two weeks ago --
  uses a list of 16 passwords as a secondary way to infect computers.

Yet another RISK of bad password choices.  And before PGN says it:
Deloder attacked de losers!  B-)

David J. Aronson, Software Engineer for hire in Washington DC area.
See http://destined.to/program/ for online resume, references, etc.


Profile of a virus writer

<"NewsScan" <newsscan@newsscan.com>>
Tue, 18 Mar 2003 09:03:24 -0700

According to the UK's Sophos, one of the world's largest antivirus
companies, about 1,000 viruses are created every month, and in almost all
cases the perpetrators are computer-obsessed males between the ages of 14
and 34. "They have a chronic lack of girlfriends, are usually socially
inadequate and are drawn compulsively to write self-replicating codes. It's
a form of original graffiti to them," says Sophos CEO Jan Hruska. Virus
writers tend to explore known bugs in existing software or look for
vulnerabilities in new versions in order to create and spread their
infections, and Hruska notes that the next target for the virus writing
community could be Microsoft's .Net platform for Web services. To boost the
impact of their creations, virus writers also tend to share information to
create variants of the same infection, such as the infamous Klez worm, which
has been among the world's most prolific viruses in the last year.
[Reuters/CNet News.com 18 Mar 2003; NewsScan Daily, 18 March 2003]
  http://news.com.com/2100-1002-993023.html?tag=fd_top


Search engines making sensitive information easy to locate

<Richard Moore <rich@westpoint.ltd.uk>>
Thu, 13 Mar 2003 11:53:32 +0000

I work for a company that performs vulnerability assessment using tools such
as the nessus security scanner. Yesterday I was searching for more
information about one of the security holes in a report, and I came across
someone else's report with the same hole. Looking a bit further, I noticed
that searching for 'nessus report' I got a dozens of pages back.  Some of
the pages returned were sample reports from companies like ours (or the
nessus site itself), but others were from people who had left their reports
in location visible to search engines.

The danger here is obvious - do you really want to make a list of all your
security holes visible to the world? Things are actually even worse than
they first appear too, because engines like google cache the pages, so even
if you delete them people can still access the information.


Benetton clothes to include tracking chip

<Monty Solomon <monty@roscom.com>>
Wed, 12 Mar 2003 13:06:07 -0500

Clothes sold at Benetton stores will soon contain microchip transmitters
that allow the Italian retailer to track its garments from their point of
manufacture to the moment they're sold in any of its 5,000 shops.
  [Source: Associated Press, 11 Mar 2003]
  http://news.lycos.com/news/story.asp?section=Business&storyId=673780

Tag, You're It: What Your Clothes Say About You
Clothing designer Benetton plans to weave radio frequency ID chips
into its garment tags. While Benetton is poised to save money by
tracking the clothes with RFID, it could also mean a loss of
customers' privacy.   [Source: Elisa Batista, wired.com]
  http://www.wired.com/news/wireless/0,1382,58006,00.html


CASPIAN calls for immediate worldwide boycott of Benetton

<Monty Solomon <monty@roscom.com>>
Wed, 12 Mar 2003 18:53:33 -0500

An American consumer privacy group has called for an immediate, worldwide
nboycott of Benetton, following disclosures that the company has placed
identification and tracking devices into its clothing products.  CASPIAN
(Consumers Against Supermarket Privacy Invasion and Numbering) announced
that it will oppose Benetton's plans to place Radio Frequency Identification
(RFID) chips into clothing labels intended for the consumer market.

For additional information, see Phillips/Benetton press release:
  http://biz.yahoo.com/bw/030311/115697_1.htm

Forbes article illustrating remote inventorying of shoppers' clothing
(As reproduced on Alien Technology's Web site)
  http://www.alientechnology.com/news/The_Internet_of_Things.htm

CASPIAN overview of privacy concerns associated with RFID technology:
  http://www.nocards.org/AutoID/overview.shtml


Federated network identity

<"brian-h seborg" <brian.seborg@telispark.com>>
Tue, 18 Mar 2003 17:57:55 -0500

One of the "Holy Grails" of networking has long been the concept of single
sign-on.  For those not familiar with the term, it refers to the ability for
a user to authenticate once to some authentication server and then to have
access to many different systems in a secure but unimpeded manner.  Kerberos
includes the concept of single sign-on within its framework as does DCE.
X.509 certificates and associated PKI also hold forth the concept of
single-sign-on through the use of certificates that are issued by a trusted
issuing authority, and can be presented as a means of authenticating the
user.  More recently, Microsoft has come up with a Web-based identity in
their Microsoft Passport standard.  A competing standard to Microsoft
Passport is that being put forth by the Liberty Alliance Project.  In all of
the standards above, there is the concept of users having a set of
attributes or credentials associated with their authenticated identity.
These credentials can be used to determine whether or not users should have
access to certain systems, and, further, to determine the level of access.
The attributes can also include such things as age verification, credit card
information, address, etc. that could be used with on-line merchants for
purchase or other purposes.  However, as this concept has matured, there is
now the thought that users would be able to have (Microsoft Passport and
Liberty Alliance), a single federated network identity, and that the user
would have some control over what information he would be willing to share
with different communities (e.g., a user could have a work profile
associated with his ID that would provide certain salient information to
work related computers and a home profile that would provide potentially
different information to computers that the same user accesses from home).
The ability to have such a federated network identity would, if it could be
administered in a secure fashion, enable any Web site or Web application
that accepted the federated identity (once authenticated by an identity
provider) to have a clear idea of whom they are dealing with.  In other
words, rather than Joe Smith connecting to a site as Joe Smith, JSmith,
Jdog, etc. and appearing to be one or multiple users, he would be Joe Smith
always.  Now, as you can imagine, the ability of Joe to be able to determine
what information he wants to share with different sites would be important.
For example, Joe may be quite willing to share his full name, mailing
address, and credit card number with an on-line store, but only be willing
to share his age with a porn site that may be allowing free tours to
prospective customers who can provide age verification.  The problem is,
that, as we have seen with the misuse of social security numbers by other
than the Social Security Administration in the U.S., we can expect to see
sites demanding more information than they legitimately need in order to
allow access.  So, this places the burden (and perhaps rightly so) on the
individual to make sure that he makes good choices as to what information he
is willing to share with each community.  This may be quite a bit to expect
of someone who is not technically savvy enough to understand the
ramifications of these choices.  This risk can be mitigated to an extent by
looking to a third party to help an individual reduce some of this risk.
For example, if I connect to a portal provided by my bank and the bank
performs the identity verification to ensure that I am who I say I am, then
I may assume that any vendors that look to the bank to verify my identity
and associated credentials have, themselves, been scrutinized by the bank
and determined to be trustworthy at least to the extent that they have
privacy policies that are similar to that of the banks so that information
provided to them is treated with the same care as that provided to the bank.
Unfortunately, I think this is contrary to reality.  So, the point is, be
very suspicious of Microsoft Passport or of the Liberty Alliance Project, or
any other group who is trying to create the equivalent of a universal ID.
While I think that they have some very good ideas, and are trying to create
a federated network identity with the best of intentions, there is much that
is not being covered that may call for not only study by consumer groups,
but for legislation and additional infrastructure.  Even then, I would still
want to be able to anonymously surf the Web the same way I anonymously enter
stores at the mall, knowing that I can walk in, look around and even
purchase something without the store knowing who I am.


Re: Computer crashes threaten hospital operations (Solomon, R-22.62)

<Jonathan Kamens <jik@kamens.brookline.ma.us>>
Tue, 11 Mar 2003 13:04:29 -0500

As far as I could tell from the article, what happened at Beth Israel
Deaconess in November 2002 [year fixed in RISKS ARCHIVES. PGN-oops] was not
a "computer crash."  That's one of the terms that journalists use when they
either don't understand what went wrong or are trying to dumb it down for
the semi-literate masses.

Quoting from *The Boston Globe* article, "The computer crash at Beth Israel
occurred when a research [sic] flooded the network with large quantities of
data, causing the strained system to slow drastically.... networks must have
sufficient bandwidth and modern routing systems, and should allow for
portions to be shut down without the system becoming disabled."

In other words, the network didn't have enough bandwidth to handle the
traffic being thrown at it, there was no infrastructure in place for
limiting the amount of bandwidth used by individual computers or
applications, there was no infrastructure in place for monitoring bandwidth
in real-time to detect anomalies and trace them to their sources, and there
was no infrastructure in place for reactively locating high bandwidth
consumers.

For all of these things to be missing in a mission-critical network is
shocking (and that's why the articles about it were written), but not
terribly surprising.  Beth Israel Deaconess has been in dire financial
straits for several years; I doubt they pay high enough salaries to attract
the best and brightest IT professionals, people who understand the
importance of constructing a robust network and how to go about doing it.  I
also doubt they've budgeted enough money to be able to afford the proper
infrastructure; perhaps, after last November's outage, that will change.


Re: Monster electricity bill (RISKS-22.61-62)

<gingrich <gingrich@bigpond.net.au>>
Fri, 14 Mar 2003 07:43:05 +1100

I can comment in some detail about a similar situation.

I worked for what was the State Electricity Commission of Victoria
(Australia) in the early 1990s.  A new customer billing system was
implemented which, among other things included a "special" interface for
senior billing officers in each local office that bypassed most of the
sanity checks in the system, and allowed the sanity checks to be
over-ridden, if necessary.

The electricity meter was replaced for a customer and this was not correctly
recorded.  Then when the next, regularly scheduled meter reading was entered
into the system, it looked as if the meter had gone from a low value to an
even lower value.  This resulted in an enormous bill, a significant
percentage (about 10%, from memory) of total sales for the district.

The system asked for confirmation, but the operator (an ordinary data entry
clerk, not the supervisor) thoughtlessly confirmed the entry.  Since it was
"special" mode the transaction was confirmed.

The customer received a bill for several million Australian dollars.

The risks?

Creating data entry screens with the power to gratuitously over-ride the
sanity checks for the data that are built into the system.

Is this really a direct problem with the computer system?  I'm not sure.  It
is definitely a problem with the interaction between the computer system and
the organisational systems designed to use the computer system.

The real lesson is that there is a need to recognise that computer systems
do not exist in isolation, and that giving the power to over-ride the system
to ordinary users is a bad idea.

Don Gingrich   gingrich@cs.rmit.edu.au    School of CSIT, RMIT Melbourne, Au

  [New meters and wrap-arounds are of course old problems.  For example, see
  RISKS-7.40 and RISKS-12.16.  PGN]


Human protocol failure

<"Dawn Cohen" <COHEND@wyeth.com>>
Tue, 11 Mar 2003 16:40:21 -0500

I guess I'm a bit influenced by RISKS, and observed a disturbing
(human-based) protocol failure last week in my daughter's after school care.
I think it would come down to one of those system-boundary things.

On a normal day, my daughter goes to school on the school bus, but after
school, she attends an after school program physically located in the
school, but administered by the local YWCA.  I pick her up from the school.
Transfer of daughter among three parties: school bus -> school -> YWCA.

Last week, we had one of those days that we were being threatened with a
snowstorm (early news reports promised 3-6 inches by afternoon), and this
being New Jersey, everyone panicked.

Because there was ultimately no serious accumulation of snow, everything was
business-as-usual for most of the day.  However, at 2:00, I got a phone call
from the after school program, indicating that the program would not be
available that afternoon, due to the snow, and that I should pick my
daughter up.  I arrived at the school at 3:00, and was herded into the
cafeteria, to wait with the other anxious and disoriented after school
program parents, for the 3:10 dismissal.  At 3:10, children began filtering
into the cafeteria, and parents collected their children and left.  10
minutes later, half a dozen parents remained, and we were informed by a
teacher that all of the non-bus children had been released.  She said that
the Y had told the school to put all the after school program children onto
the school bus.  At that point, I ran out, and with some difficulty located
the school bus that my daughter comes to school on (which had luckily not
pulled out, yet), and pulled her off the bus.

System risks here:
1) The school accepts YWCA's word that it has informed all parents of
   the closure of the after-school program (what if the Y had not been able
   to reach me?)
2) The school accepts YWCA's word that it has informed all of the
   parents of  what action will be taken (this failed in my case)
3) The school releases my child to a 3rd party (the school bus) that does
   not have sufficient information for the situation.  The school has
   emergency contacts for me and alternative pick-up people, as does the
   afterschool program.  The school bus department does not have this
   information.  It was not clear what the bus driver would have done if I
   had not rescued my daughter.  I asked him what would have happened if he
   tried to drop my daughter at the house, with nobody at home, and it seems
   like their protocol is the most sensible: if there is no one to receive a
   child, radio the transportation department.  The department then calls
   the home (the only number they have for me), and if no one answers, take
   the child back to school.  If I want my neighbors to take my child, I
   have to supply a note to this effect.

I don't know.  Maybe I'm just refusing to take off my RISKS glasses, here,
but I think there's problems in the protocol with all three parties.  I've
communicated with the Principal, who refuses to accept any responsibility.
She asserts that the whole fault is with the Y, who should have informed me
that my daughter would be sent home on the bus.  Personally, from reading
RISKS, I think it is unacceptable to assume that communication was sent
correctly to all parents.  The school explicitly demands at the beginning of
the year a list of acceptable people to release my child to, and the school
bus is not on this list.  I guess a better protocol would require the Y to
provide minimal coverage at the school until such time as any parents who
have not been successfully contacted arrive.  Also, I guess I'm going to
have to suggest that the transportation department keep my emergency
contacts list, too.


Announcement: The Workshop on Rapid Malcode: WORM

<"Robert K. Cunningham" <rkc@ll.mit.edu>>
Tue, 18 Mar 2003 17:00:38 -0500

             The Workshop on Rapid Malcode (WORM)
             Workshop held in association with the
  10th ACM Conference on Computer and Communications Security,
              October 27th, 2003   Washington D.C.
                      Call for Papers

In the last several years, Internet-wide infectious epidemics have emerged
as one of the leading threats to information security and service
availability.  The vehicle for these outbreaks, malicious codes called
"worms", leverage the combination of software monocultures and the
uncontrolled Internet communication model to quickly compromise large
numbers of hosts.  Current operational practices have not been able to
manage these threats effectively and the research community is only now
beginning to address this area. The goal of this workshop is to bring
together ideas, understanding and experience bearing on the worm problem
from a wide range of communities including academia, industry and the
government.  We are soliciting papers from researchers and practitioners on
subjects including, but not limited to:

    Modeling and analysis of propagation dynamics
    Automatic detection, characterization, and prediction
    Analysis of worm construction, current & future
    Propagation strategies (fast & obvious vs slow and stealthy)
    Reactive countermeasures
    Proactive defenses
    Threat assessment
    Forensic methods of attribution
    Significant operational experiences

Paper submissions due 1 Jul 2003; submission instructions will appear at
  http://pisa.ucsd.edu/worm03/

General Chair: Stuart Staniford, Silicon Defense
Publicity Chair: Robert Cunningham, MIT Lincoln Lab
Program Committee Chair: Stefan Savage, UC San Diego
Program Committee Members: Robert Cunningham, MIT Lincoln Lab;
  Anup Ghosh, DARPA; David Moore, CAIDA/UC San Diego;
  Carey Nachenberg, Symantec; Vern Paxson, ICIR/LBL;
  Phil Porras, SRI; Jeff Rowe, UC Davis; Mike Skroch, Sandia;
  Stuart Staniford, Silicon Defense; Don Towsley, UMassAmherst

Dr. Robert K. Cunningham, Information Systems Technology Group
MIT Lincoln Laboratory  http://www.ll.mit.edu/IST/

Please report problems with the web pages to the maintainer

Top