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 49

Weds 15 January 2003

Contents

Computer sabotage against Venezuela oil?
David Wagner
Brace for onslaught of new viruses
NewsScan
Y2K+3 bug in Networker
William D. Colburn
Smut hits 'Army Newswatch'
Monty Solomon
How to vote for your favorite California quarter design
Fred Cohen
Hong Kong gym pulls plug on camera cell phones
Monty Solomon
Amazon not checking for sensible values
Jeremy Epstein
Google Search cached a password protected page?
Colin Sutton
Misuse of HTML comments causes missed comments
Alexander Dupuy
Biometric lunch lady
Richard Akerman
Re: PGP.COM cannot handle sales to some US residents
Stephan Somogyi
REVIEW: "CISSP for Dummies", Lawrence Miller/Peter Gregory
Rob Slade
REVIEW: "Information Security Policies, Procedures, and Standards", Thomas R. Peltier
Rob Slade
Info on RISKS (comp.risks)

Computer sabotage against Venezuela oil?

<David Wagner <daw@cs.berkeley.edu>>
Tue, 14 Jan 2003 21:13:37 -0800 (PST)

*Oil Daily* quoted Ali Rodriguez (head of Venezuela's state oil company):

  "[...] we have suffered many acts of sabotage at the terminals, the
   refiners, and even to some well-heads in Lake Maracaibo.  There were
   even instances of computer hacking which did a lot of damage since
   much of the operation is centrally controlled by computer."
   [Source: *Oil Daily*, vol 53, no 9, 14 Jan 2003]

Does anyone know anything more?


Brace for onslaught of new viruses

<"NewsScan" <newsscan@newsscan.com>>
Tue, 14 Jan 2003 08:48:22 -0700

Computer users will be plagued with a host of new viruses this year,
particularly worms deployed into instant messaging systems, predicts a
senior technology consultant with UK-based Sophos. "Virus writers are most
interested in creating the next super Windows worm, spread by e-mail or
instant messaging, as these mass-mailing viruses carry the greatest impact,"
says Graham Cluley. "We expect more executable e-mail-aware worms this year,
while more viruses are written which use instant messaging services." Sophos
also expects to see an increase in the number of so-called "Backdoor
Trojans," which can open up holes in operating systems so that crackers can
control them from a remote location. Windows users are particularly at risk,
as nine out of 10 of last year's top viruses were spread via e-mail on
Windows platforms, with the most prolific being the Klez worm. So far, PDAs
and mobile phones have remained largely free of virus problems, says
Cluley. "There is no indication yet that we will see an avalanche of new
viruses affecting mobile devices -- virus writers are not interested in
targeting the mobile phone until it becomes more developed and has a bigger,
common platform." [Reuters 14 Jan 2003; NewsScan Daily, 14 Jan 2003]
  http://shorl.com/fikunumubodri


Y2K+3 bug in Networker

<"William D. Colburn (Schlake)" <wcolburn@nmt.edu>>
Fri, 10 Jan 2003 11:19:43 -0700

We currently use NetWorker 6.0.1.Build.174 Turbo/15 as our backup
software.

The first Saturday of the new year (4 Jan 2003) it appeared that only one
tape was correctly labeled, and that data was appended to previous dumps on
some recycled tapes.  Now that the second Saturday is upon us, we watched
today's tape labelling and saw that Networker is labelling the first tape
correctly, but then reverting to the label scheme for the backups on 5 Jan
2002.

The risk of this error is almost social engineering.  Tomorrow's tapes
should be labelled Full.2003.01.11.a, .b, etc.  Some tapes labelled as
Full.2002.01. will soon be eligible for recycling.  We recycle by hand, and
if the person doing it didn't know these were mislabled then we could lose
recent backups.

William Colburn, Computer Center, New Mexico Institute of Mining and Technology
wcolburn@nmt.edu   http://www.nmt.edu/tcc/   http://www.nmt.edu/~wcolburn


Smut hits 'Army Newswatch'

<"monty solomon" <monty@roscom.com>>
Thu, 9 Jan 2003 22:34:06 -0500

The town of Webster, NY, is investigating what might have caused a program
about the U.S. Army on its local cable access channel to be interrupted by
about 20 minutes of explicit gay porn.  [Source: John Kohlstrand, *Rochester
Democrat and Chronicle*, 9 Jan 2003; PGN-ed]
  http://www.rochesterdandc.com/news/0108story8_news.shtml


How to vote for your favorite California quarter design

<Fred Cohen <fc@all.net>>
Tue, 14 Jan 2003 15:55:01 -0800 (PST)

The forms at:
	http://www.caquarter.ca.gov/

Allow anyone to vote on the appearance of the California quarter.  According
to my daughter you can vote as many times or as often as you want.  Ther
site uses a simple 'FORM NAME="COIN" METHOD="post"' method with parameters
like 'NAME="QUARTER" VALUE="16"' (my vote) as parameters.

It is indeed trivial to send in a few hundred votes per second from a modem,
thousands from a DSL line or cable modem, and of course if you can afford an
OC48, you can probably vote billions of times a day.

I don't think they will use this as their sole decision process, but when it
comes to Internet voting, I think this is a shining example of how artistic
license can mint a winner.

Fred Cohen  http://all.net/  fc@all.net  fc@unhca.com  tel/fax: 925-454-0171
Fred Cohen & Associates	- University of New Haven - Security Posture

  [However, the new quarters are apparently revitalizing coin collecting,
  which had grown dormant of late.  So, maybe there are also plans to have a
  similar popularity vote for your favorite state quarter, fully expecting
  to report votes of 435,829,662,554 for the Rhode Island coin in hopes that
  it will sell more.  PGN]


Hong Kong gym pulls plug on camera cell phones

<Monty Solomon <monty@roscom.com>>
Wed, 15 Jan 2003 10:58:34 -0500

Warning: use of camera-equipped mobile phones could be hazardous to your
health.  That's the message going out from at least one chain of health
clubs in Hong Kong, where a new generation of cell phones that can take and
transmit video and still photos is raising concerns over a new crop of
privacy-related issues.  Physical, which operates nine gyms in the former
British colony, recently posted signs in its Hong Kong facilities forbidding
the use of mobile phones in locker rooms.  [Source: Reuters item by Doug
Young, 14 Jan 2003]
  http://news.lycos.com/news/story.asp?section=OddNews&storyId=623861


Amazon not checking for sensible values

<"Jeremy Epstein" <jepstein@webmethods.com>>
Fri, 10 Jan 2003 11:41:29 -0500

There are lots of bugs due to not checking for sensible values, and I found
this one amusing.  Bret Hartman sent me a copy of his new book "Mastering
Web Services Security", and I wanted to recommend it to a friend.  So I went
to Amazon, which tells me "Availability: This title will be released on
December 31, 1969. You may order it now and we will ship it to you when it
arrives."

Seems to be missing a sanity check regarding the "to be published date"
being in the future.


Google Search cached a password protected page?

<"Colin Sutton" <colin@sutton.wow.aust.com>>
Sun, 12 Jan 2003 22:30:46 +1100

Searching for SQL help on google
http://www.google.com.au/search?q=%22ANSI+SQL+CASE%22&hl=en&lr=&ie=UTF-8&oe=
UTF-8&start=10&sa=N I found a link to a password controlled site
http://rec.schoolsnet.syd.catholic.edu.au/ , but the page and hundreds more
on that site are visible in the google cache.

Can a supposedly secure site be cached on google for anyone to see if a
person with access permissions uses google to search it?

I don't have access to any such sites, or I'd try it.

Or, perhaps the password protection was added after the site was protected.

Either way, there's a risk.


Misuse of HTML comments causes missed comments

<Alexander Dupuy <dupuy@cs.columbia.edu>>
Sat, 11 Jan 2003 01:31:08 -0500

My boss sent me (using some version of Outlook) some important mail with
comments on several included nested messages, all of which in multipart
alternative text and html.  Unfortunately, my browser chose the html, which
contained lots of microsoft stuff embedded in HTML comments.

Apparently there's a bug in whatever version of Outlook, and the microsoft
gunk sometimes omits a closing <![endif]-->, which isn't a big deal for the
microsoft HTML renderer, but is a big deal for other HTML renderers, since
it leaves the HTML comment open, including all of the message text up to the
end of the next HTML comment, just before the body text of the first
"Original Message".

For microsoft HTML renderers, the missing bit makes no difference, as it is
displaying the data even though it is in a comment (I guess it recovers okay
from the missing endif), but for people with other HTML renderers, the
comment is completely invisible.  I only saw the included messages, but none
of my boss's comments (the non-HTML kind).

I only realized that this had happened when one of my co-workers (who uses
emacs to read his e-mail) replied to the message, and included some text
that I hadn't seen.


Biometric lunch lady

<Richard Akerman <rakerman@bigfoot.com>>
Thu, 09 Jan 2003 21:53:02 -0500

In what seems like a remarkably elaborate solution to a simple problem,
"students will be charged for their lunches with a retina scanning device"
at a school in England starting in September.
  http://www.salon.com/tech/wire/2003/01/09/uk_school/
This brings a whole new meaning to "stealing someone's lunch money".


Re: PGP.COM cannot handle sales to some US residents (Kabay, R-22.46)

<Stephan Somogyi <ssomogyi@pgp.com>>
Sun, 5 Jan 2003 23:04:29 -0800

It distresses me acutely that Dr Kabay did not have the most friction-free
purchasing experience possible -- my inner Ferengi does so hate it when we
hinder a customer from exchanging money for one of our fine products -- and
while the above provides all the necessary clues, Dr Kabay unfortunately
does not draw the correct conclusions.

As he points out, Dr Kabay is using his StarBand account. This is a
satellite-based ISP with a footprint that covers the continental US as well
as quite a bit of geography surrounding it.
<http://www.starband.com/whatis/> is a helpful source for determining the
service's reach.

I am not a professional cartographer, yet it appears to my layman's
understanding of geography that one of the above explicitly named seven
countries quite likely finds itself within a satellite footprint that
extends out to the US Virgin Islands.

Since PGP Corp is located in California, we are subject to US laws.  These
include the export restrictions overseen by the Department of Commerce in
the form of its BIS, the Bureau of Industry and Security, formerly known as
the BXA, the Bureau of Export Administration.

If we guess right about a given satellite ISP user, we generate about 40
bucks in revenue. If we guess wrong, we're looking at $10K+ in fines, and
are further subject to administrative actions up to and including "no more
export to anywhere for you, ever."

As risk assessment goes, this is not a hand-wringer.

>The customer service agent was very nice and obviously embarrassed
>about this situation and admitted that there are no measures in
>place for dealing with such a technical glitch.

There is no technical glitch. Evidence clearly suggests that our customer
service agent may not have had all the facts available to her at the time Dr
Kabay called, but the system functioned as designed.

We "fail closed" rather than "fail open" based on having examined the risks
inherent in delivering export-controlled software to a world-wide
audience. If we cannot ascertain your whereabouts with a reasonable degree
of certainty (and StarBand doesn't provide that information as part of the
reverse-resolved IP address), we will not deliver software to you and expose
ourselves to some rather substantial liability.

>1) Check the IP address BEFORE the user fills out all the forms and
>the credit card gets debited.

I'm delighted that Dr Kabay suggests this course of action, since it
provides the opportunity to expose another risk that we've had to take into
consideration. It is not feasible to perform such a check before the user
fills out the form since it opens us up to a financial denial of service
attack.

Performing the necessary checks to determine a buyer's suitability for sale
costs PGP Corp money per check. The process involved is not a simple matter
of reverse-resolving an IP address. If we were to perform the requisite
check prior to accepting the credit card, it would be a trivial exercise to
launch an attack that's quite expensive to us.

>2) Send the user a CD-ROM to the US address listed in the order.

We are currently investigating this as an alternate delivery method.

>3) Ask the user for strong evidence that they are in fact living in
>the US:  e.g., [...]

I cannot imagine that Dr Kabay meant this suggestion seriously. As a
non-governmental organization, we have no ability to verify a given driver's
license's validity or the whereabouts of its holder, nor are we able to
divine what an "appropriate US fax machine" is.

Moreover, it strikes me as ironic to suggest that a company such as PGP
Corp, which is committed to security and privacy, put itself in a position
where it has to make a determination about what constitutes an authoritative
means of location identification, and then verify it.

>b) ask for other corroborating evidence such as a US address
>listing in university or corporate Web sites.

Had Dr Kabay attempted to purchase and download our software from a
university or corporation with an unambiguous location, I daresay he
would've succeeded.

>RISKS of assuming your automated system is perfect: you lose sales.

We assume no such thing. Our system did, however, operate precisely
as it was supposed to in this instance.

RISKS of not erring on the side of caution when it comes to the law: gaining
the concerted attentions of the BIS.

Stephan Somogyi, Director of Products, PGP Corporation


REVIEW: "CISSP for Dummies", Lawrence Miller/Peter Gregory

<Rob Slade <rslade@sprint.ca>>
Tue, 10 Dec 2002 08:02:38 -0800

BKCISPDM.RVW   20021029

"CISSP for Dummies", Lawrence Miller/Peter Gregory, 2002,
0-7645-1670-1, U$39.99/C$59.99/UK#29.95
%A   Lawrence Miller
%A   Peter Gregory peter.gregory@hartgregorygroup.com
%C   5353 Dundas Street West, 4th Floor, Etobicoke, ON   M9B 6H8
%D   2002
%G   0-7645-1670-1
%I   John Wiley & Sons, Inc.
%O   U$39.99/C$59.99/UK#29.95 416-236-4433 fax: 416-236-4448
%O  http://www.amazon.com/exec/obidos/ASIN/0764516701/robsladesinterne
%P   408 p. + CD-ROM
%T   "CISSP for Dummies"

A 'cheat sheet' is bound into the front of the book.  It offers some general
advice for taking the CISSP (Certified Information Systems Security
Professional) exam, the most useful aspect of which is to prepare.  Most of
the tips are vague, such as the suggestion to budget your time, or review
CISSP resources, without any information about what factors should be
considered in time management or where to find resources.  Some tips are
overly specific, such as the recommendation that you bring a big bottle of
water.  (Yes, six hours is a long time for the exam, and, yes, you may need
refreshment.  The tip does not mention that proctors vary in rigour when
applying the exam regulations, and may not allow bottles of water at the
test tables.  Besides which, only one person may be excused from the room at
any one time.)

Part one reviews the CISSP exam itself.  At the beginning of chapter one,
the authors point out that some CISSP study guides are too hard, and some
CISSP study guides are too soft, but this book is just right.  Then it moves
on to information about (ISC)^2 (the International Information Systems
Security Certification Consortium), arrangements for the exam, and some
study tips.  The material is more up-to-date than in other CISSP study
guides, but the text is badly written, duplicating content and repeating
itself, possibly because the structure and organization is weak.  The
suggestions and information are reasonable, although occasionally
questionable: the recommendations for study guides and practice exams are
rather weak.  Chapter two briefly lists the ten domains of the common body
of knowledge (CBK), and is really only an expanded table of contents for the
chapters in the next section.

Part two describes the ten domains in detail.  Chapter three covers most of
access control, but unevenly.  Given the constraints that the authors
themselves mention (the CISSP CBK is a mile wide and an inch deep), too much
space is devoted to a simplistic set of password choice rules, an excellent
(but, in this situation, overlong) review of Kerberos, and a number of jokes
which are not going to help candidates remember important points, and may
very well confuse the issues.  Some material is problematic, such as the
discussion of security "domains" that follows the Microsoft networking model
rather than the Bell-LaPadula derived structure that the CBK requires, and a
baffling non-explanation of the lattice model.  (There are also a number of
perplexing inclusions, such as a cross-reference to cryptography in the
introduction to single sign-on systems.)  Telecommunications and network
security is presented in chapter four.  The authors have used the OSI (Open
Systems Interconnection) model to structure the discussion of various
technologies: an interesting concept, but one which is flawed by the fact
that a number of topics are placed in the wrong level.  (Media access and
packet switching, for example, are listed in the data link layer, rather
than the physical and network layers, respectively.)  There are also
problematic references to "native" PPP (Point-to-Point Protocol) encryption,
and an assertion that ICMP (Internet *Control* Message Protocol) packets are
not required for network operations.  The basics of security management are
covered in chapter five, but very tersely.  The major standards are not
listed here: the Common Criteria is mentioned briefly in chapter eight
(security architecture) but British Standard 7799/ISO (International
Standards Organization) 17799 is not listed at all.  The set of roles and
responsibilities is short and risk analysis terms are not well defined.
This must be considered a serious weakness in the book, since security
management is very important in the CISSP exam.  Application development is
dealt with briefly and poorly: again, this is an area where many CISSP
candidates do need extra help, and they won't get it here.  System
development methods are not discussed at all, and the malware section is
full of errors.  (Each chapter lists a set of books for extra research: I
should note that neither of the virus books listed at the ISC2 site appear
on the list for this chapter.  In fact, the bibliography is rather short
overall: Krutz and Vines "The CISSP Prep Guide" (cf.  BKCISPPG.RVW) which is
not much better than the current work, is listed in every set.)  There are
also odd inclusions from other domains, such as almost a full page devoted
to the SYN flood attack, which was adequately explained in a paragraph in
chapter four.  The material on cryptography, in chapter seven, lists all the
terms and technologies, but has poor or non-existent explanations,
mathematical errors, and the authors obviously do *not* understand S-boxes.
(The process described would not allow for decryption.)  There is too much
text about CPUs (Central Processing Units), and too little on distributed
systems, formal models, and the various evaluation criteria in chapter
eight's review of security architecture.  Operations security, in chapter
nine, seems to be a collection of random topics, with a fair concentration
on audit logs.  Chapter ten's overview of Business Continuity Planning (BCP)
is not bad, although a bit shy on details.  (The vital topic of backups, for
example, is mentioned only long enough to say that you should have one, and
the various types, with varying strengths and weaknesses, are not discussed
at all.)  Law, investigation, and ethics is reasonable, although the list of
specific privacy laws is probably not too helpful (and I rather suspect that
the authors got taken in by the "Desert Storm Virus" myth).  Most of the
material on physical security, in chapter twelve, appears to have been
copied from some other source without much understanding: the sections on
visibility, capacitance sensors, and UPSes (Uninterruptible Power Supplies)
are among those that contain errors or seem to miss the major points.

Part three is the usual "dummies" "part of tens."  Chapter thirteen relists
the ten domains.  (Didn't we do this already?)  Ten other security
certifications are recorded in chapter fourteen.  Web sites are given in
chapter fifteen: three are actually useful.  The cheat sheet and chapter one
are reprised in sixteen and seventeen.  One of the books listed in chapter
eighteen ("Security Engineering," by Ross Anderson, cf. BKSECENG.RVW) would
be very useful for exam candidates.

Sample test questions are a big part of every CISSP study book (in the case
of Peltier and Howard's "The Total CISSP Exam Prep Book," in fact, the
*only* part).  This book has both its own set of questions, and a set from
the Boson exams.  As I have said elsewhere, the Boson exams are not
necessarily wrong, but they are far too simplistic to be considered adequate
preparation for the CISSP exam, and the answer guides are completely tied to
"Secured Computing" (cf. BKSCDCMP.RVW).  If any set of questions are
simpler, and therefore less useful, than the Boson set, they are the ones
listed in this book.  And, like the Boson collection, the answers are
completely self-referential.

Like Andress' "CISSP Exam Cram" (cf. BKCISPEC.RVW), this text does sometimes
simply list the terminology, although Miller and Gregory are somewhat more
complete and do provide greater explanations of the domains themselves.  It
would be hard to make a distinction between this volume and "Secured
Computing": Miller and Gregory provide *some* outside references but Endorf
makes fewer errors.  As previously noted, Krutz and Vines do not give the
reader much in the way of explanatory material, but they do cover the
domains more comprehensively than the current work.  Harris' "CISSP
All-in-One Certification Exam Guide" is, as noted (cf. BKCISPA1.RVW), the
one guide that might get you through the CISSP exam, albeit not necessarily
with high marks: Miller and Gregory might get you through, but only if you
stood a pretty good chance without the volume.

copyright Robert M. Slade, 2002   BKCISPDM.RVW   20021029
rslade@vcn.bc.ca  rslade@sprint.ca  slade@victoria.tc.ca p1@canada.com
http://victoria.tc.ca/techrev    or    http://sun.soci.niu.edu/~rslade


REVIEW: "Information Security Policies, Procedures, and Standards",

<Rob Slade <rslade@sprint.ca>>
Wed, 4 Dec 2002 08:43:30 -0800
  Thomas R. Peltier

BKISPPAS.RVW   20020923

"Information Security Policies, Procedures, and Standards", Thomas R.
Peltier, 2002, 0-8493-1137-3
%A   Thomas R. Peltier
%C   920 Mercer Street, Windsor, ON   N9A 7C2
%D   2002
%G   0-8493-1137-3
%I   Auerbach Publications
%O   U$69.95 +1-800-950-1216 auerbach@wgl.com orders@crcpress.com
%O  http://www.amazon.com/exec/obidos/ASIN/0849311373/robsladesinterne
%P   297 p.
%T   "Information Security Policies, Procedures, and Standards"

Chapter one provides vague meanderings about information protection
fundamentals.  The author's opinion about how to write is given in
chapter two.  In the ultimate triumph of style over substance, this
drafting advice is given before any examination of actual policy
development.  Chapter three defines policy and some related topics
with lots of verbiage and overly lengthy examples.  There are lots of
sample mission statements in chapter four, although it is not really
apparent why we are talking about this particular topic.  The
structure of chapter five, dealing with standards, is very confused,
and the purpose of the examples given is unclear.  (There is also an
extremely odd assertion that standards, which are by definition rigid,
must be "flexible.")  We are given more writing advice, supposedly in
aid of procedures, in chapter six.  Chapter seven talks about
information classification for a few paragraphs and then lays out a
thirty page example.  Random security thoughts and banal training
ideas make up the security awareness program in chapter eight.
Generic project management advice is in chapter nine.  Chapter ten
contains suggested topics for a security policy.  What the book said
is repeated in chapter eleven.

The appendices include a very short sample policy, and a policy
development checklist.

Barman's "Writing Information Security Policies" (cf. BKWRINSP.RVW)
provides far better advice on both the process and the topics to be
covered in creating a security policy.  Even "Information Security
Policies Made Easy" (cf. BKISPME.RVW) is better, for all that people
tend to misuse it.  Peltier's book provides little of use to the
harried security manager.

copyright Robert M. Slade, 2002   BKISPPAS.RVW   20020923
rslade@vcn.bc.ca  rslade@sprint.ca  slade@victoria.tc.ca p1@canada.com
http://victoria.tc.ca/techrev    or    http://sun.soci.niu.edu/~rslade

Please report problems with the web pages to the maintainer

Top