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 26 Issue 28

Wednesday 12 January 2011

Contents

Lots of computer risks related items in the news lately
Robert Schaefer
Storage on the cloud is not the only risk
Geoffrey Brent
Risk of coffee in the cockpit
Nick Brown
Car Theft by Antenna
Robert Schaefer
Estonia prep's for CyberWar
PGN
"SMS of Death" Could Crash Many Mobile Phones
Robert Schaefer
Microsoft investigates 'phantom' Windows Phone 7 data
Robert Schaefer
Risks of E-health records
PGN
Cornell digester overflow due to "programming error"
Doug Elrod
FedEx: we can't deliver your package because the computer says so
Jonathan Kamens
WSJ: Russia, China go open-source—distrust U.S. commercial software
Lauren Weinstein
Privacy-Enhanced Mobile Data Storage and Self-Destruct Mechanisms
Lauren Weinstein
Stock Market Cheating Risk?
Gene Wirchenko
I am stupid, and it has cost me
Paul Robinson
Re: RISKS of reusing ID numbers
Jonathan Kamens
Re: WikiLeaks
Patrick Gustafson
Re: Radiation Machines Overdosing Again
Kevin Fu
Alexandre Peshansky
Info on RISKS (comp.risks)

Lots of computer risks related items in the news lately.

Robert Schaefer <rps@haystack.mit.edu>
Tue, 11 Jan 2011 07:47:36 -0500

Networks in hospitals may become 'regulated medical devices'.
http://www.computerworld.com/s/article/9203761/FDA_eyes_regulation_of_wireless_networks_at_clinics_hospitals?taxonomyId=132&pageNumber=1


News
FDA eyes regulation of wireless networks at clinics, hospitals
Networks could become 'regulated medical devices' because of data they carry

By Lucas Mearian

robert schaefer, Atmospheric Sciences Group, MIT Haystack Observatory
Westford, MA 01886 781-981-5767  http://www.haystack.mit.edu


Storage on the cloud is not the only risk

Geoffrey Brent <gpbrent@optusnet.com.au>
Sat, 01 Jan 2011 14:34:13 +1100

Jeremy Epstein (RISKS 26.27) commented on the risks of cloud storage. I
suspect we're also going to see an increase in risks associated with
outsourcing of processing.

For instance, if you're in a delivery business and want to optimise your
vehicle routing, you will need to perform a large number of driving-distance
calculations. Calculating driving distance is a moderately complicated
problem requiring use of road-net data, so you have three options. You can
spend time and money writing your own calculator and keeping it updated - or
you can spend several thousand dollars on a stand-alone commercial product -
or you can write a simple function call that queries Google Maps for the
data you need.

>From discussion with others, I get the impression this is an easy risk to
 miss. A lot of people who'd be horrified by the idea of storing
 confidential addresses on a Google cloud service or sending them via GMail
 would be quite blase about feeding them into a Google Maps query. My guess
 is that it's about intent: when we use email or storage services, it's with
 the goal of putting data somewhere, so we tend to ask "where is this
 going?" But with a query, our intent is to *get* data, making it easier to
 forget that the flow of information is two-way.


Risk of coffee in the cockpit

"Nick Brown" <Nick.BROWN@coe.int>
Wed, 5 Jan 2011 09:49:11 +0100

A United Airlines transatlantic flight from Chicago to Frankfurt was
diverted to Toronto after the pilot dumped a cup of coffee on the plane's
communication's equipment. This triggered a series of emergency codes,
including one for a hijacking.

I don't want to speculate what might have happened if the coffee had also
taken out the plane's voice communication with the tower.  However, all is
well: as long as the person in the pilot's seat can convince ATC that they
are not, in fact, a hijacker, that alarm can be safely ignored.  (Just
kidding: I'm sure that there are all kinds of protocols in place before the
Air Force shoots you down.)

http://edition.cnn.com/2011/TRAVEL/01/05/canada.flight.diverted/index.html


Car Theft by Antenna

Robert Schaefer <rps@haystack.mit.edu>
Fri, 7 Jan 2011 09:42:11 -0500

Researchers beat automatic locking and ignition systems.
No key required: A researcher shows how an attacker could start a car using
an antenna. A signal from the car is transmitted to a computerized key,
which is tricked into enabling the engine ignition.  ETH Zurich
http://www.technologyreview.com/computing/27037/?a=f


Estonia prep's for CyberWar

"Peter G. Neumann" <neumann@csl.sri.com>
Wed, 5 Jan 2011 7:29:23 PST

  [Thanks to Phil Porras.  PGN]

Volunteer Cyber Army Emerges In Estonia

In April 2007, the Baltic republic of Estonia became the first country in
the world to experience cyberwar. Government, financial and media computer
networks were paralyzed by a series of attacks, which authorities ultimately
concluded originated in Russia.

In the years since that cyberassault, Estonia has distinguished itself once
again: Now it is a model for how a country might defend itself during a
cyberwar. The responsibility would fall to a force of programmers, computer
scientists and software engineers who make up a Cyber Defense League, a
volunteer organization that in wartime would function under a unified
military command.
http://www.npr.org/2011/01/04/132634099/in-estonia-volunteer-cyber-army-defends-nation?sc=tw&cc=share


"SMS of Death" Could Crash Many Mobile Phones

Robert Schaefer <rps@haystack.mit.edu>
Fri, 7 Jan 2011 11:54:44 -0500

Phones don't have to be smart to be vulnerable.
http://www.technologyreview.com/communications/27021/


Microsoft investigates 'phantom' Windows Phone 7 data

Robert Schaefer <rps@haystack.mit.edu>
Tue, 11 Jan 2011 07:44:43 -0500

Microsoft has told BBC News that it is investigating why some handsets
running its Windows Phone 7 software are sending and receiving "phantom
data".  Several net forums detail complaints from people that say their
phones are automatically eating into their monthly data plans without their
knowledge.  Some have complained that their phone sends "between 30 and 50MB
of data" every day; an amount that would eat into a 1GB allowance in 20
days.

http://www.bbc.co.uk/news/technology-12152517

robert schaefer, Atmospheric Sciences Group, MIT Haystack Observatory
Westford, MA 01886 781-981-5767  http://www.haystack.mit.edu


Risks of E-health records

"Peter G. Neumann" <neumann@csl.sri.com>
Tue, 4 Jan 2011 16:59:34 PST

  [Thanks to dkross]

"Most importantly, the electronic medical record affects how we think. The
system encourages fragmented documentation, with different aspects of a
patient's condition secreted in unconnected fields, so it's much harder to
keep a global synthesis of the patient in mind."
http://well.blogs.nytimes.com/2010/12/30/the-doctor-vs-the-computer/


Cornell digester overflow due to "programming error"

dre1@cornell.edu
Tue, 4 Jan 2011 12:38:36 -0800 (PST)

Cornell University's animal-remains digester recently discharged brine-like
"effluent" into local streams.  Causes of the incident were found to
include:

- Remote re-programming by a manufacturer's representative who assumed his
- changes would be automatically reset the next day.
- Lack of communication between the re-programmer and local operating staff.
- Failure modes leading to effluent reaching ventilation ducts (!), and
- thereby leaving the building.

Some details are at http://www.news.cornell.edu/stories/Dec10/VetDigester.html.


FedEx: we can't deliver your package because the computer says so

Jonathan Kamens <jik@kamens.us>
Sun, 02 Jan 2011 00:58:26 -0500

A GIGO story that RISKS readers may find interesting...

A couple months ago, FedEx failed to deliver a package to my house three
times in a row because no one was home.

Using the door tag they left for us, I contacted them and asked them to hold
the package at a FedEx location near my office.

I went there with a picture ID and asked for the package. The told me, "That
package isn't addressed to [my house number] [my street]. It's addressed to
[my house number+1] [my street], and the name on it doesn't match your ID."

Yes, that's right, the FedEx driver had attempted to deliver the package to
the wrong house three times in a row. I pointed out the erroneous delivery
attempts and told him he'd better make sure the package was redelivered to
the correct address. He said he'd take care of it.

Meanwhile, the intended recipient of the package called FedEx to find out
what had happened to her package. They informed her that they had made three
failed delivery attempts, followed by an attempt to hold the package for
pickup which failed because the person who came to get it did not have a
valid picture ID, so they had no choice but to return the package to its
sender. The fact that the delivery attempts were made to the wrong house was
not mentioned.

She argued at length with several people, telling each of them that there
had been no delivery attempts to her house, that she had not asked for the
package to be held anywhere, and that if she had, she would have had a valid
picture ID, so clearly whoever asked for the package to be held wasn't
her. They all insisted they could not deliver the package, but she did
finally convince one of them to give her the phone number of the person who
made the hold request (i.e., me).

We are passing acquaintances, so she recognized my phone number. She called
me and asked what was up, and I told her what had happened.

Armed with that additional information, she called FedEx back. She spoke to
several people who were apologetic and sympathetic and yet at the same time
insistent that there was nothing they could do. Since the computer said
there were three failed delivery attempts and a failed hold attempt, they
simply could not redeliver the package. No one seemed to have any idea how
to tell the computer that the delivery attempts had been to the wrong
address. Either that, or they were unwilling to do so (didn't want to get a
driver in trouble? didn't want to damage the performance statistics for
their location?).

Finally, she got a supervisor to agree to deliver the package. He brought it
over to her house in his own car, out of uniform, late that night, i.e.,
completely outside the system. One cannot help but wonder what FedEx's
computer says about the package now. Alas, I neglected to save the tracking
number so I can't find out.

A possibly related fact is that when the package was finally delivered, it
was all beaten up and had been broken open and patched up in transit,
despite the fact that it was in a brand new box when sent.

I contacted FedEx's executive complaints office by email several weeks ago
and asked them to comment on how this happened and whether supervisors
delivering packages late at night in their own cars is standard operating
procedure for FedEx. I got no response.


WSJ: Russia, China go open-source—distrust U.S. commercial software

Lauren Weinstein <lauren@vortex.com>
Sat, 8 Jan 2011 21:05:25 -0800

  From Network Neutrality Squad: nnsquad@nnsquad.org

WSJ: Russia, China go open-source—distrust U.S. commercial software
http://on.wsj.com/dPoeOU  (WSJ)

My favorite quotes from the piece:

 "At the end of 2010, the "open-source" software movement, whose
  activists tend to be fringe academics and ponytailed computer geeks,
  found an unusual ally: the Russian government." ...
 "But just a few weeks before Mr. Putin publicly endorsed open-source
  software, FBI Director Robert Mueller toured Silicon Valley's leading
  companies to ask their CEOs to build back doors into their software,
  making it easier for American law enforcement and intelligence
  gathering agencies to eavesdrop on online conversations."

And a bit of dialogue that I've frequently quoted over the years from
one of the most brilliant dark "comedy" films of all time:

Russian Spy:  "Are you trying to tell me that every phone
               in the country is tapped?"

American Spy: "That's what's in my head..."

Russian Spy:  "But Don!  This is AMERICA... not RUSSIA!"

                    --- "The President's Analyst" (1967)

Lauren Weinstein (lauren@vortex.com) http://www.vortex.com/lauren
+1 (818) 225-2800 http://www.pfir.org  http://www.nnsquad.org
Global Coalition for Transparent Internet Performance: http://www.gctip.org


Privacy-Enhanced Mobile Data Storage and Self-Destruct Mechanisms

Lauren Weinstein <lauren@vortex.com>
Wed, 5 Jan 2011 17:25:44 -0800

            Urgent Call for Privacy-Enhanced Mobile Data Storage
                       and Self-Destruct Mechanisms
               http://lauren.vortex.com/archive/000797.html

Greetings.  Once upon a time—not so very long ago—an individual
arrested by law enforcement, or subjected to search at border custom
checkpoints, would typically be carrying little more of interest than
clothing, a purse or wallet containing limited sundry items, and more
recently a very simple cell phone.

But now many of us carry powerful computing devices that frequently contain
immense volumes of personal and business data—laptops, smartphones,
tablets, flash memory thumb drives, and soon other yet to be imagined
marvels.  While it is increasingly possible to store data only in the cloud
for download or streaming on demand, many users still need to maintain
significantly large amounts of data on their local devices due to data
access speed requirements, or to assure data availability when remote data
connections are not available.

Governments in general and law enforcement in particular are increasingly
taking the view that their detailed inspections of mobile devices, and the
masses of data that they frequently contain, are no different in kind than a
simple search of a suspect's or traveler's pockets.

Now the California Supreme Court has alarmingly ruled that arrested
suspects' phones—and by extension any other devices on their person or in
their vehicles at the time of their arrest—can be comprehensively
searched in detail.  This includes all contained data, without the need for
a search warrant: "Photos, address book, Web browsing history, data stored
in apps (including social media apps), voicemail messages, search history,
chat logs, and more."  ( http://bit.ly/gdUj6K [CNN] )

While this ruling is not without conflict vis-a-vis some rulings in other
states, and may ultimately be decided by the U.S. Supreme Court, it still
appears on its face to represent an enormous overreaching of law enforcement
in a highly inappropriate manner.

As I mentioned above, international travelers have long faced the risk of
U.S. Customs not only inspecting the data on their laptops or other
computers upon reentry to the U.S., but of having those devices arbitrarily
confiscated for detailed inspection, data copying, and other intrusive
investigations for prolonged periods of time.

If the framers of the U.S. Constitution had been able to anticipate that
individuals would one day carry such vast quantities of information
representing virtually the sum totals of their business and personal lives,
it is likely that the Fourth Amendment prohibiting unreasonable searches and
seizures would have been written in ways that even more explicitly
prohibited "high-tech" data device strip searches.

It's very important to remember that this is not about protecting criminal
behavior—we're talking about the protection and preservation of
fundamental constitutional rights, that are now being eroded by
opportunistic overreaching on the part of authorities (whether for laudable
motives in any given case or not).  Nor can we confidently assume that all
future governments will even be as "benign" as our own at any given time --
encroachments on privacy rights by government are fundamentally dangerous,
especially for innocent, law-abiding citizens.

Fortunately, we do have the means at hand to restore some sense of balance
regarding the privacy of our personal, mobile data devices.

The powerful combination of local device storage, increasingly fast
"persistent" data connections, cloud-based data repositories, high-grade
encryption, and associated technologies, can provide the foundation for an
open-source framework to provide privacy-enhanced mobile data storage and
data "self-destruction" systems to help return "search and seizure" closer
to the concept that the Founding Fathers had in mind.

So, I'm now making this urgent call for broad cooperation in the development
of *open-source* systems and environments that would include at least the
following initial attributes:

- Provide for the "continuous and automatic" backing up of *all* mobile
  device data (as desired) in secure off-device locations.  Such locations
  could include cloud-based services and/or locally controlled
  (e.g. business or home) computer systems and data arrays.  Note that under
  current laws the precise physical location of data greatly impacts the
  required mechanisms for government inspection or seizure of that data.
  Mobile devices (certainly in California for now) are pretty much an open
  book after the new Supreme Court ruling.  Various groups are working
  toward trying to achieve harmonization of laws to provide the equivalent
  of locally-hosted data privacy protections for cloud-based data, but
  battles in this regard are still ahead.  Also, the ability of authorities
  to try compel the provision of data decryption keys and related
  information varies depending on situations, jurisdictions, and other
  factors.

- Users should be able to optionally specify degrees of data security
  desired on a per-item basis.  For data without significant privacy-related
  concerns, mobile device data self-destruct mechanisms could be flagged to
  bypass that specific data (e.g. specific files, databases, etc.) under
  particular usage scenarios.  Individual data items could also be flagged
  for various degrees of off-device data repository security—unencrypted
  (e.g. publicly shared data), encrypted, or various combinations as
  appropriate.

- All communications between mobile devices and remote data repositories
  would be encrypted.

- Mobile device data self-destruct mechanisms would be designed to enable
  ease of use in routine, unusual, and emergency situations for selected or
  full data.  For example, a traveler about to enter U.S.  customs could use
  a routine activation sequence to "cleanse" sensitive business data from a
  mobile device, then restore it completely (restoration priority at the
  control of the user) afterward.  In unusual or emergency situations, data
  self-destruct activation may be through a unique device key sequence or
  carefully confirmed voice command sequence.  Sequences to delete
  off-device stored backup data in remote repositories, and methodologies
  for remote triggering of mobile device data self-destruct (including both
  manually triggered and "tamper triggered" sequences, would likely be
  commensurately more complex to avoid undesired data loss, depending on the
  level of backup data chosen and available.

- Self-destruct/deletion procedures for stored data (both locally stored on
  mobile devices and to the greatest extent possible on remote repository
  backup data systems), would be designed to offer varying levels of
  resistance to forensic deleted data reconstruction, as specified by users
  for particular data and usage scenarios.

I hope that's enough to get the ball rolling.  It's very important that such
concepts be implemented in an open-source environment, and that strong,
high-grade encryption be used throughout the framework wherever encryption
is employed.

Again, this is most definitely not about protecting illegal activities or
criminals.  The goal is to protect us all—and our completely legal
personal, business, and other data—from unreasonable acts by those
entities who are now leveraging our advanced mobile data devices to a level
of intrusion into our lives that is simply not in keeping with our
fundamental rights and liberties.

While I do have my own very preliminary, somewhat specific implementation
concepts relating to this project, I'm very much inviting all comers and all
ideas.  In terms of practical project goals, I would encourage the
development of these principles into exploratory code as rapidly as
possible, across a wide array of mobile platforms and supporting backup
repository system environments.

Linux, Windows, and Android are currently available to me in various
incarnations.  Google's Cr-48 Chrome notebook would be another obvious
implementation target platform that I would like to explore early on for the
project, though unfortunately I do not have one of those units in hand.

I am not a routine user of the Apple ecosystem, so developers comfortable in
the Mac/iPhone world are definitely needed as well, plus Blackberry,
Symbian, and any other common mobile platforms.

Please let me know if you're interested in participating.  Any and all
comments, questions, criticisms, and ideas are of course welcome.

Thanks all.  Be seeing you.

Lauren Weinstein (lauren@vortex.com) http://www.vortex.com/lauren
+1 (818) 225-2800  Twitter: https://twitter.com/laurenweinstein
Blog: http://lauren.vortex.com   Google Buzz: http://bit.ly/lauren-buzz


Stock Market Cheating Risk?

Gene Wirchenko <genew@ocis.net>
Thu, 06 Jan 2011 10:41:43 -0800

InfoWorld Home / The Industry Standard / Tech's Bottom Line
January 06, 2011
Hackers find new way to cheat on Wall Street—to everyone's peril
'Side-channel' attack on high-frequency trading networks could net a
hacker millions of dollars in just seconds—and leave everyone else
that much poorer
http://www.infoworld.com/d/the-industry-standard/hackers-find-new-way-cheat-wall-street-everyones-peril-699

Opening paragraphs:

High-frequency trading networks, which complete stock market transactions in
microseconds, are vulnerable to manipulation by hackers who can inject tiny
amounts of latency into them. By doing so, they can subtly change the course
of trading and pocket profits of millions of dollars in just a few seconds,
says Rony Kay, a former IBM research fellow and founder of a cPacket
Networks, a Silicon Valley firm that develops chips and technologies for
network monitoring and traffic analysis.

Kay, an Israeli-born computer scientist and one-time Intel engineering
manager, says the root of the problem is the increasing speed of networks;
as they get faster and faster, our ability to actually understand events
taking place within them isn't keeping up.  Network monitoring technology
can detect perturbations in network traffic happening in milliseconds, but
when changes occur in microseconds, they're not visible, he says.

cPacket has developed a proof of concept showing that these "side-channel"
attacks can be used to create tiny delays in the transmission of market data
and trades. By manipulating specific trading activities by several
microseconds, an attacker could gain unfair trading advantage. And because
the operation occurs outside the range of monitoring technology, it would
remain invisible. "We believe that such techniques pose a substantial risk
of creating unfair trading, if used by the wrong people," Kay says.


I am stupid, and it has cost me

Paul Robinson <paul@paul-robinson.us>
Mon, 10 Jan 2011 13:34:44 -0800 (PST)

I was working on my new computer, which, because I have lots of data files
on my previous computer, I have both plugged in and the old one runs
headless.  By having them networked I can use SMB (Windows networking) to
share files from one computer to another, I can basically use my previous
computer as a file server.

I have two external hard drives, one which is a 250 GB and the other is a 1
TB.  The 250 is about 1/2 full as I moved my music collection over to it
because the 160 GB drive in my old main computer was running low on space.

A few weeks ago, while moving things around, El Stupido (yours truly) moves
his computer without removing the external hard drives, which were sitting
on top.  Maybe I can plead disability; I am working from a wheelchair and
sometimes I forget to watch where I am.  Maybe I can plead insanity, I
should have been thinking.  But it doesn't matter what I plead; the universe
doesn't provide an appeals process when you make a mistake.  They both slide
off and fall on the floor, about 2-3 feet.  But it was enough.

They don't work anymore.  I plug them in, one makes noise and the other just
clicks.  I wasn't really using the terabyte drive, but my entire music
collection was on the 250GB.  Did I have a backup?  Well, yes and no; I had
an older backup on DVDs but that was in the folder that got lost when I
moved from my previous apartment.  El Stupido didn't have a backup and kept
the drive where it could get damaged if he wasn't careful.

I have no idea how many files I lost; I can estimate my entire music
collection, probably 3,000 mp3s, WAVs, and OGGs, maybe 40 GB of files.  As
the drive is essentially suffering from physical damage, it would have to be
fixed in a recovery facility, probably by disassembling in a clean room.
Cost: about $1100.00.

I had deleted the current files off my other computer because I needed the
space.  I have - or I hope I have - some of the original files on my other
computer in the closet.

I can't believe I was this stupid.  You know, it's funny; it's pointed out
that the typical value of a hard drive is only a fraction of the value of
the data.  Or the cost of the recovery if lost.

All I can say is, make sure you have a backup, or that you have two.  A 1
terabyte USB hard drive used to be about $100; if you look carefully you can
now find 2 terabyte drives for that.  If you don't need that much space,
external USB hard drive prices are as low as 10c per gigabyte, which makes
hard drive and DVD pricing close, with DVDs at 4c/GB, the only problem being
DVDs being small compared to the tremendous amounts of data we typically
have around.  Small discs have an advantage that they are easier to store
than hard drives, so that can be solved with larger discs.  Blu-Ray burners
can be gotten for about $120 or so; 25GB Blu-Ray discs sell for about $20 in
10-pack spindles.

Like Jacob Morley in "A Christmas Carol" I carry a ponderous chain of lost
data, and I can only offer my example in a hope that you will not repeat it.
"Otherwise you cannot hope to escape my fate."

Of course now, if you don't recognize the risks from my story, I have some
oceanfront property in Nevada I'd like to sell you...


Re: RISKS of reusing ID numbers (RISKS-26.27)

Jonathan Kamens <jik@kamens.us>
Wed, 05 Jan 2011 06:52:25 -0500

I did not "miss the point" of Geoff Kuenning's incident report, but he
clearly missed mine.

Even the best-designed system for assigning tracking numbers can result
in number reuse in situations which are either unavoidable or not the
fault of the carrier. Kuenning's assertion that the incident he observed
is proof of an "indefensible" design is not the only possible, or even
necessarily the most likely, explanation for the evidence he presented.

His assertion that, "It is quite plausible that two shipments could be
confused in a situation that would be much more difficult to spot, and
much more expensive to resolve," is also not supported by the limited
evidence he provided. It is easy to imagine a system in which the
carrier has ensured that even when a tracking number is reused
relatively quickly, it is sufficiently separated by time and distance
from the previous use that there is no ambiguity. In fact, even with a
tracking number namespace as "small" as nine digits, given the number of
possible origins and destinations for shipments, the odds of two
shipments with the same tracking number having similar beginning and end
points are vanishingly small. Finally, even though the carrier's Web
site only shows the origin and destination cities, their internal data
obviously shows the precise origin and destination addresses as well as
the sender and recipient names, so in the extremely unlikely event of
ambiguity, the resolution is as simple as calling the carrier on the
phone, which is hardly "expensive."

Kuenning's assertion that "a longer ID number is cheap to generate and
process" is also not supported by his evidence. It could be, for
example, that the carrier involved had far more growth and success than
its business plan anticipated, and by the time it was carrying enough
traffic that its tracking numbers were clearly too short, there was an
entrenched, worldwide infrastructure using the nine-digit number which
would be extremely expensive to retool (and nevertheless, for all we
know the carrier may very well be in the process of retooling it).
Sometimes technical concerns must give way to business concerns.

In summary, Kuenning has presented a single incident which lends itself
to many different, equally plausible explanations, and his insistence
that the most RISKSy one is the only possible one is incorrect.


Re: WikiLeaks (RISKS-26.27)

"Patrick Gustafson" <PGustafs@wms.com>
Tue, 4 Jan 2011 11:13:36 -0600

Not a risk per se, but in your ending comment you mention "ubiquitous losses
of personal privacy".  I'm not sure that anybody should take it for granted
that anything done over electronic communication is private.  Maybe I'm
being alarmist, but I worked on CALEA, and lawful intercept systems in GSM
networks, and understand that privacy doesn't really exist. And with the
internet, it is even easier to intercept traffic without detection (sure
some users can, but not a significant percentage). =20

But then that's why I like the Risks Digest so much.
Paranoidly yours,

patrick gustafson | principal systems engineer, wms  773-961-1083


Re: Radiation Machines Overdosing Again (Ladkin, RISKS-26.26)

Kevin Fu <kevinfu@cs.umass.edu>
Sat, 8 Jan 2011 17:31:43 -0500

Stanley F. Quayle writes in RISKS 26(27):
> It might work differently elsewhere, but here in the USA, the "standard"
> is to sue everyone: the manufacturer, the airline, the FAA, the pilots,
> all the way down to the mechanics that last touched the airplane.

Barry Gold writes in RISKS 26(27):
> You can bet that every one of the patients
> killed (or injured) by an accidental overdose has received compensation, or
> soon will.

Counter-intuitively, medical device manufacturers in the United States are
largely immune to lawsuits from patients.  In Riegel v. Medtronic Inc., the
U.S. Supreme Court ruled 8-to-1 [1] that "a medical-device manufacturer
cannot be sued under state law by patients alleging harm from a device that
received marketing approval from the Food and Drug Administration (FDA)
[2]."

Thus, patients have little legal recourse if harmed or killed by an approved
device.  The Medical Device Safety Act of 2009 intended to restore some
recourse to patients without enabling frivolous lawsuits, but was stuck in
Waxman's Energy and Commerce Subcommittee on Health when I last attended the
hearings.  It is unknown how the 112th Congress will weigh patient safety
vs. corporate interests.  It's a hard problem to solve because of the need
to balance these competing interests.  See [3,4] for related discussion.
I'd be interested to hearing how other countries tackle this challenge.

[1] "Justices Shield Medical Devices From Lawsuits" by Linda Greenhouse.  In
   The New York Times, February 21, 2008.
[2] "The Medical Device Safety Act of 2009" by Curfman et al.  In New
   England Journal of Medicine, March 18, 2009.
[3] "Semper Fidelis—Consumer Protection for Patients with Implanted
   Medical Devices" by William H. Maisel, In New England Journal of
   Medicine, March 6, 2008.
[4] "Trustworthy Medical Device Software" by Kevin Fu.  In Public Health
   Effectiveness of the FDA 510(k) Clearance Process: Measuring Postmarket
   Performance and Other Select Topics: Workshop Report, Institute of
   Medicine, National Academies Press, preprint available as Appendix D of
   http://www.nap.edu/catalog.php?record_id=13020 or
   http://www.iom.edu/Activities/PublicHealth/510KProcess/2010-JUL-28.aspx

Kevin Fu, Assistant Professor, Computer Science Department, University of
Massachusetts Amherst 616-594-0385 http://www.cs.umass.edu/~kevinfu/


Re: Radiation Machines Overdosing Again (RISKS-26.26)

Alexandre Peshansky <Alex.Peshansky@einstein.yu.edu>
Mon, 3 Jan 2011 19:13:09 +0000

In RISKS-26.27 Barry Gold commented on the above article and quoted an
anecdote from decades back about how ingenious fools are just using bigger
hammers to defeat fool-proof design.

This, in turn, reminded me of a story I lived through.

In my previous life in the USSR I worked in airspace industry, on the
factory that produced fuel pumps for jet engines.  Now, "fuel pump" may be a
misnomer, as these were essentially mechanical computers, calculating fuel
volume as a function of pilot's control, air temperature and engine speed
(yes, a four-dimensional problem).  So, as you can imagine, there were a lot
of very precisely machined parts in these pumps.  In fact, the precision was
beyond manufacturing ability, so the pumps were assembled using so called
"selective assembly".  The parts were sorted by size after manufacturing and
carefully laid out in assembly trays in matching pairs.

Once the engine of a fighter jet failed in flight, the plane crashed, and
the subsequent investigation nailed seized pump.

The investigators traced the pump to us, and any RISKS reader can guess what
happened: an assembly worker used the wrong (unmatched) part.  As it did not
fit, he used a hammer.  As this particular fluidic circuit was working in a
very specific conditions, it was not caught during the testing...

Computer risks relevance: the testing was semi-automated (I was working on
computerization of testing equipment); my algorithms were, of course, based
on existing testing modes and would not caught this either.  After the
accident, I totally revised the approach and redesigned algorithms to go
through all possible logic branches.  (By this time USSR was going down the
drain, and I left the factory before new testing stands were implemented -
AFAIK, they never were).

Alexandre Peshansky, Snr. Bioinformatics Analyst, ICTR/RIC
Albert Einstein College of Medicine of Yeshiva University (718) 430-2440

Please report problems with the web pages to the maintainer

Top