TITLE>LISTWATCH: Security List Items

News Items from Security-Related Mailing Lists

by Mary Ellen Zurko, Open Group Research Institute (zurko@osf.org)

This issue's highlights are from privacy, www-security, e$pam, risks, tbtf, ietf-tls and dcsb.

A search engine called " Magellan Internet Guide" (http://www.mckinley.com) offers an option called Search Voyeur (exactly what I felt like), which displays a continually updated list of "20 randomly selected real-time searches that users like you are now performing on Magellan" (like me?). I was treated to several selections including "nude celbs", "how can meet tory aikman", "high school bands", "europiccola", "floral design projects", and "sexual harassment policy". Searches can clearly include worrisome information like personal names, and it's not clear from the home page (if you're just intent on issuing a search) that your search parameters could be broadcast to the Net.

A bug in Netscape 3.0 (fixed in 3.01) allowed a web page with Javascript to send email (along with the configured email address) without warning the user. Discussion again brought up the need for more levels of defense (beyond disabling Javascript) such as signed applets.

There was a lot of discussion about the significance of the cryptanalytic DES (and 3DES) attack announced by Biham and Shamir, which is a theoretical attack based on inducing faults via physical access (such as with microwave radiation). The concern is mostly about smartcards, and the attack was compared to other sorts of fraud that is tolerated with other kinds of payment mechanisms (cash, check, credit cards). An experiment verifying the theoretical results would raise the level of concern somewhat. They announced follow-up research on discovering secret keys even when the operation of the cryptosystem is unknown (such as with Skipjack).[See news item on this topic with pointers to related papers -- CEL]

Another challenge to ITAR based on freedom of speech is being raised by Peter Junger, a law professor at Case Western Reserve University. He argues that he can't even discuss the legality of ITAR in the presence of non-US students as the law is written ( http://sun.soci.niu.edu/~cudigest/CUDS8/cud870).

Discussion on the IETF's Transaction Level Security (TLS) working group centered on overall strategy and authentication methods. There is pressure to produce a document, and a cleaned-up version of SSL v3.0 would be easy and useful for future SSL developers. Proposals have been advanced for adding secret key authentication and Kerberos to TLS 1.0. Proponents cite practical deployment issues; opponents worry that weakness in chosen passwords or secret key schemes could damage the percieved strenght of TLS.

Microsoft distributed a CD-ROM with a document infected with the WAZZU.A Word Macro virus at an exhibition in Switzerland. Even when MS officials were made aware of this virus, the CD-ROM continued to be distributed, and the document was available for downloading from a Web site.

A Dallas judge has issued what is thought to be the first injunction delivered via email and usenet. "It's our position that under Texas state law whenever a person has knowledge of an order, that is sufficient notice," one of the attorneys in case said. For those of you who subscribe, the URL is http://www.nytimes.com/library/cyber/week/1017harass.html.

Privacy Assured, which is a pilot program of the Electronic Frontier Foundation's eTrust project, will post its blue PA logo on Web sites that adhere to its standards. These standards include: not knowingly listing information about individuals that has not been volunteered for publication; not allowing reverse searches to determine individuals' names from e-mail addresses, phone numbers or other information; releasing only aggregated usage statistics, not individual information; and giving individuals the option to delete personal information from lists.

Adam Shostack shared information derived from his real-world experience with code reviews. They find security & reliability bugs at about one per 20-50 lines of code, which is dropping to closer to one per hundred as he distributes copies of code review guidelines he wrote ( http://www.homeport.org/~adam/review.html). Reviewing superficially takes about an hour for 500-1000 lines of commented code. A deep review to find tricky problems can take much longer. They've found that a review team of fewer than 4 people is less effective at finding problems. Reviewing more than about 2000 lines of code (2-3 hours) in a day can cause burn-out.

The latest Clinton administration proposal for "key recovery" caused a lot of discussion. The proposal would still include communication keys, which are unnecessary for the recovery of data that is stored encrypted and only useful for the monitoring of communications. The main issue is privacy of these communications. It is still not clear what key recovery mechanisms the market would produce, since the government will have to approve of key recovery plans before allowing 56 bit encryption to be exported. Nonetheless, it's the first proposal with a carrot component, and many companies have announced that they tend to comply. Apple, often cheered in privacy advocate circles, has announced they will comply (along with IBM, Digital, RSA, and others), while Microsoft, often reviled in the same circles, is holding out (along with Netscape).

Mitsubishi Electric Corp. said it has released the design principles and a sample program for its proprietary data encryption algorithm MISTY ( http://www.melco.co.jp/rd_home/new/crypt_e.html). Proprietary encryption algorithms have a poor reputation in the cryptographic community, since they are often written by cryptographers with little experience, and many cryptographic algorithms have flaws that are only discovered after extensive and lengthy (years) peer review. I was not able to find the location of this information.

Rumor has it that one or more spoof sites were able to take advantage of the syntax issues of URLs. Bob Dole announced a site at something like dole-kemp96.com at the end of one of the debates. Some variant of trying with and without the hyphen and using org instead of com got you to a spoof site. I can no longer find any spoofs, and both dolekemp96.com and dolekemp96.org look fairly legitimate.

Dallas Semiconductor announced something called an iButton, which can be used for authentication. There were concerns raised over its random number generator and PIN protection. A co-worker showed me one embedded in a plastic ring, and it definately evoked a "Captain Crunch Decoder Ring" feel, which made me want one.

Jeff Schiller, IETF Security Area Director, formally recommended to the IPSEC working group that the mandatory key management protocol for IPv6 be ISAKMP/Oakley, and that SKIP be optional.