- cypherpunks-legacy - lists.cpunks.org

RC4 article in Saturday (Sept 17) New York Times
by Matt Blaze 17 Dec '03

17 Dec '03

John Markoff has a piece on the RC4 betrayal in the Business section of the Saturday NY Times (page 37), "A secret computer code is out -- Key to data security appears on internet". Not much that hasn't already been said here or on sci.crypt, but there was an interesting quote from Jim Bidzos that suggested that one of the conditions RSADSI agreed to in order to get approval of 40 bit RC4 for export in shrink-wrap software included keeping the algorithm confidential. Bidzos speculated that the NSA could revoke RC4's export status as a result of the disclosure. Also, the piece reports that "The RC4 formula was first circulated on Tuesday to a specialized computer network mailing list of computer researchers who oppose the Government's stringent controls on data encryption technology. The mailing list, which has thousands of computer users around the world, is known as Cypherpunks, and the mailings usually consist of highly technical discussions of data encryption technology." I guess Markoff gets Eric Blossom's moderated version of the list :-) -matt

3 2

Nice little privacy threat...
by L. Todd Masco 17 Dec '03

17 Dec '03

Just a little tidbit for the privacy-paranoid among us. I live in NYC, while my father, H. L. Masco, MD, lives in Florida. My dad's never lived in NYC, and I've never forwarded mail from his address to mine... in fact, I haven't forwarded anything from his address since before I left Pittsburgh. I just got a mailing from Harry & David that included a pre-printed list of folks from "my" last year's orders so I could repeat order this year. They were my father's orders from last year. Masco isn't that common a surname, so I have to wonder how often this is going to happen to people with more common surnames. -- L. Todd Masco | "A man would simply have to be as mad as a hatter, to try and cactus(a)bb.com | change the world with a plastic platter." - Todd Rundgren

1 0

Re: Netscape competence
by L. McCarthy 17 Dec '03

17 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- - -----BEGIN PGP SIGNED MESSAGE----- Marc Andreessen writes: > In article <199412091814.NAA07757(a)hermes.bwh.harvard.edu>, adam(a)bwh.harvard.edu (Adam Shostack) wrote: > > > It my personal feeling that Netscape doesn't have the right > > talent mix to develop secure software. For example, they may well get > > the RSA parts right, and then store the passphrase in a text file, > > 'for ease of use.' > > My goodness, that's a bit malicious and unsubstantiated, isn't it? Excuse me, but Adam Shostack didn't write that paragraph you quoted. Amanda Walker <amanda(a)intercon.com> did, as I recall. BTW, welcome to the list.... - - -L. Futplex McCarthy - -----BEGIN PGP SIGNATURE----- Version: 2.6.1 iQCVAwUBLuuJZ2f7YYibNzjpAQERXQP/SJnxzt7tckazsBBtfQrKA3c08RLjYUs9 1xK3b/WqqFvmCDyMpwQg5yLzEk7uq7SFp725FMP/jZ8ZUnvpQGBxPdEaFOxpjlHU cPZOb1R1OzLQ2m0023H6LyBM3WMlCd7Nah+zOqaoTbQ1kofE9hZ2gXEfljurUy+8 7scWCfdrZfw= =sIUK - -----END PGP SIGNATURE----- - --- [This message has been signed by an auto-signing service. A valid signature means only that it has been received at the address corresponding to the signature and forwarded.] -----BEGIN PGP SIGNATURE----- Version: 2.6.2 Comment: Gratis auto-signing service iQBFAwUBLuuKryoZzwIn1bdtAQF/jgF/SPy47fk1vvZgyXl8UBBDuASGJn+COQC5 84E+ICfuxxvxR4GauaaLLwCPdzH4l1DL =KuB4 -----END PGP SIGNATURE-----

1 0

Further comments on Netscape et al.
by Amanda Walker 17 Dec '03

17 Dec '03

-----BEGIN PRIVACY-ENHANCED MESSAGE----- Proc-Type: 4,MIC-CLEAR Content-Domain: RFC822 Originator-Certificate: MIIB4DCCAXUCBQJBAAI3MA0GCSqGSIb3DQEBAgUAMGMxCzAJBgNVBAYTAlVTMSAw HgYDVQQKExdSU0EgRGF0YSBTZWN1cml0eSwgSW5jLjEyMDAGA1UECxMpVW5hZmZp bGlhdGVkIFVzZXIgQ2VydGlmaWNhdGlvbiBBdXRob3JpdHkwHhcNOTQwNTE2MDAw MDAwWhcNOTYwNTE2MjM1OTU5WjB6MQswCQYDVQQGEwJVUzEOMAwGA1UEERMFMjIw NzAxETAPBgNVBAgTCFZpcmdpbmlhMRAwDgYDVQQHEwdIZXJuZG9uMR4wHAYDVQQJ FBUyMDEgRWxkZW4gU3RyZWV0IKYxNzMxFjAUBgNVBAMTDUFtYW5kYSBXYWxrZXIw XDANBgkqhkiG9w0BAQEFAANLADBIAkEAsYTQHrGTWxpS22owJLNVkmBNxGW6Z8TK 5Qdsg9co9J8uyHOuqEIHIxOIDsVA7X/bsYVXxvitmc5PYKp2RuWj+QIDAQABMA0G CSqGSIb3DQEBAgUAA1YAIR3BS02rsquq3dY6sGWcieXW3HtChMtsReRhDdY/nGPz rIUYuWt087c4T4afrsG9FbouJOkAsR/5SkLI9UpXonDIUskdl4/JlRKpyLaX07RK mUyTSR== MIC-Info: RSA-MD5,RSA, lOMm69acu9EUeHUNrqC3tVCGCxCYQn24LMZIdJLsHHh+yHQUzowi3hysYfSuJEeR 15mklYr0pLdP8HQCyqv4mg== Sigh. I wasn't intending to ignite a firestorm on the list, especially on a topic that is at this point only marginally crypto-related. I bear Netscape and its staff no ill will--they have done some wonderful stuff, and I'd be more than happy to compete with them however the market ends up. I am irritated by some of how they've gone about things, but this is as much cultural as anything else--the market will decide in the end, whatever they or anyone else do. If I have offended anyone on the list or at Netscape, I'm completely willing to continue the discussion in email, comp.infosystems.www, or wherever else, but I'm going to stop sending my replies to cypherpunks at this point unless they actually have to do with cryptography or WWW security per se. Vendors do take potshots at each other from time to time; I'm sorry that I indulged myself and did so on this list instead of in a more appropriate manner. Amanda Walker InterCon Systems Corporation -----END PRIVACY-ENHANCED MESSAGE-----

1 0

Interse
by mgream＠acacia.itd.uts.edu.au 17 Dec '03

17 Dec '03

-- Matthew Gream,, M.Gream(a)uts.edu.au -- Consent Technologies, 02-821-2043.

1 0

FEDGOVT> OTA Report - Electronic Commerce - via ftp (fwd)
by Adam Shostack 17 Dec '03

17 Dec '03

Date: Tue, 31 May 1994 13:46:55 CDT Newsgroups: bit.listserv.govdoc-l From: "Dexter, Martha (Dir.,Info/Pub)" <mdexter(a)ops.ota.gov> Subject: OTA Report on Electronic Commerce ----------------------------Original message---------------------------- May 26, 1994 OTA EXAMINES FUTURE ELECTRONIC COMMERCE Lawmakers and businessmen at the turn of the century reacted only after new technologies had restructured their society. However, citizens today have an opportunity to comprehend and prepare for the radical changes taking place as the concept of a National Information Infrastructure (NII) moves from vision to reality, says the congressional Office of Technology Assessment (OTA). The Clinton Administration announced in September 1993 an initiative to promote the development of NII ... "that would create a seamless web of communications networks, computers, databases, and consumer electronics that will put vast amounts of information at users' fingertips. ...[That] can help unleash an information revolution that will change forever the way people live, work, and interact with each other." The initiative relies on the private sector to innovate and aggressively pursue the deployment of these technologies. But certain problems in the deployment of the NII will persist that only the government can address, says OTA. In testimony today before the House Committee on Science, Space, and Technology, OTA released the findings of its Report Electronic Enterprises: Looking to the Future. In the Report, OTA takes a strategic look at the development of electronic commerce, and outlines the characteristics of the infrastructure that must support it. To support U.S. businesses and to ensure a competitive economic playing field, the information infrastructure will need to be flexible and open, seamless and interoperable and evenly and broadly deployed, says OTA. The Report lays out a number of government strategies to promote a network architecture that meets these requirements. The US is in the midst of a transition created in part by advances in communication and information technologies. These developments are radically altering the US economy and changing the way that business is conducted, says OTA. Markets are expanding globally; business organizations are streamlining; what we normally think of as a firm is becoming blurred; some worker skills are becoming obsolete requiring worker retraining;and production is being carried out "just-in-time" on a flexible schedule, rather than being mass-produced. These changes are fundamental and far-reaching. OTA emphasizes that in an electronically networked economy, the design and underlying architecture of the global information infrastructure will have a major impact on national economic growth and development. If American businesses are to benefit fully from electronic commerce, says OTA, not only technology but also social and economic factors that govern the use of networking technologies need adequate attention in designing infrastructure policy. To address these factors, the government will need to look beyond the traditional role of "regulator," and consider the full range of strategies that it might pursue, says OTA. In its other various roles as broker, promoter, educator, and institution- builder, the government must establish the rules of the game and the incentive structure that will help determine private sector choices. As regulator, government will need to ensure that electronic markets are evenly deployed, open, and accessible on an equitable basis. Whether in work relationships within a firm, competition in the marketplace, or trading relations among nations, having access to information and the ability to use it strategically will be the keys to success or failure. What is different today, OTA points out, is the extent to which knowledge is now embedded in information and communication technologies. As a result, choices about the design, architecture and structure, or the rules and regulations of network technologies will be irreversible in the short- to medium-term. Requesters for the study are the Senate Committee on Commerce, Science, and Transportation, and the House Committee on Science, Space, and Technology. Copies of the 190-page Report Electronic Enterprises: Looking to the Future will be available in about three weeks from the Superintendent of Documents, U.S. Government Printing Office, and from the National Technical Information Service. To inquire about availability, call OTA at (202) 224-8996 or e-mail pubsrequest(a)ota.gov. For copies for congressional use, please call (202) 224-9241. A 4-page Summary and testimony delivered at the hearing are available electronically. (The Report will be available electronically in about a week.) To download via ftp from OTA, use the following procedures: ftp to otabbs.ota.gov (152.63.20.13). Login as anonymous. Password is your e-mail address. File is in the directory /pub/elenter. OTA is a nonpartisan analytical agency that serves the U.S Congress. Its purpose is to aid Congress in the complex and often highly technical issues that increasingly affect our society. Martha Dexter mdexter(a)ota.gov Director, Information Management (202) 228-6233 Office of Technology Assessment fax: (202) 228-6098 U.S. Congress Washington, DC 20510

1 0

Re: BofA+Netscape
by shamrock＠netcom.com 17 Dec '03

17 Dec '03

[Welcome to the list, Marc. Great to have you here!] Marc Andreessen wrote >In article <9412111647.AA23311(a)tadpole.tadpole.com>, db(a)Tadpole.COM (Doug >Barnes) wrote: > >> I'll throw in that from the perspective of someone running a server, >> their approach of requesting all graphics simultaneously over >> different sockets in the name of client performance is disastrous. >> This causes most servers to fork N times more per page, where N is >> the avg. # of graphics. > >That's just plain not true. Servers don't fork any more often >with Netscape than they do with other clients -- EVER. > >> but some have speculated that this was done deliberately in order >> sabotage server software other than their own. > >That's also just plain not true, and completely unsubstantiated. Marc, have you ever been in an empty lab with an htpd server - late at night? You can HEAR it when Mozzilla hits the site. Two or three Mozzilla users at the same time will kill your server. Unless, I understand, it uses your server software for which you charge money. Can you be surprised that there are a significant number of people out there who are wondering if your "selfless" deed of giving away your client software for free was really all that selfless? -- Lucky Green <shamrock(a)netcom.com> PGP encrypted mail preferred.

1 0

Re: Clipper in patent trouble?
by koontzd＠lrcs.loral.com 17 Dec '03

17 Dec '03

>From: Mike Ingle <MIKEINGLE(a)delphi.com> >Rich Lethin <lethin(a)toast.ai.mit.edu> wrote: >> >>I was chatting today with someone moderately well-informed about the >>clipper controversy (unlike me). He pointed out the following work to me >>by Prof. Silvio Micali at LCS on a technical scheme which can serve as a >>compromise between the needs of society for legitimate wiretaps and the >>need of individuals for strong privacy. Basically, it seems to be a >>protocol for extending a public key algorithm into a k-escrow system. This >>apparently differs from the Clipper chip in that algorithmic details are >>well publicized. >> >>The work was in Crypto '92 apparently, and an MIT lab for CS tech report >>numbered TR-579b. > >Even worse, Micali is claiming that his patent on fair cryptosystems >(#5,276,737) covers Clipper as well. In the Wall Street Journal (May 31, >1994, p. B6): > > Mr Micali, whose patent was issued in January, says his patent > covers the concept of breaking an encryption key into multiple > parts that are guaranteed to work, and are held by escrow > agents. > >It seems to me that Clipper does not guarantee that the multiple parts >will work in anywhere near the same way as his scheme does (see my book >for details); Clipper is simply a secret splitting scheme. On the other >hand, Micali filed his patent application in Apr 92, a full year before >Clipper became public. >Bruce (Schneier) One thing of note from the book on Crypto 92, is that the conference occurred in August. The paper on fair crypto systems contains references to President Clinton and Clipper, having been written or revised between May and August of 93. Hardly seems fair for something supposedly presented (and reviewed in 92). Is it just me or does this seem questionable? (Possibly being an attempt to show prior art, or perhaps being simply revised by someone close to the publication process.)

1 0

RE: NSA breaks Russian PRNGs with neural networks??
by koontzd＠lrcs.loral.com 17 Dec '03

17 Dec '03

>From: rishab(a)dxm.ernet.in > >tcmay(a)netcom.com: >> An interesting article by Seymour Hersh is cited below. It says that >> NSA had transcripts of the 1991 coup plotters (and presumably other >> Russian leaders) and that Bush passed these on to Yeltsin to warn him. > >A recent article from the Daily Telegraph, another British paper, went on abou >the possible encryption techniques used by the Russians. It described how >reused one-time pads led to the unmasking of Fuchs, the Rosenbergs, Philby >et al. Then it suggested that the method the NSA broke was based on (presumabl >weak) PRNGs, a stream cipher. It suggested that the NSA might have developed >techniques to find patterns in PRNG outputs through neural networks, or geneti >algorithms!! > >While the latter sounds like crap to me, even though I've worked with and >believe in the power of neural networks for amazing pattern recognition, unles >the PRNGs were _really_ weak, I'm skeptical. I don't think the Russians are >fools, and in these times one doesn't rely on secret weak algorithms for >crypto, not when there are publicly well known strong ones. Humint? Maybe. I can recall having seen keylists for Soviet crypto, similar (but larger) than those used for shift register based U.S. tactical crypto from the Korean War era. We used to monitor send/receive ciphertext for U.S. crypto during key changes. One handy tool was a meter, which would integrate (low pass) the data stream. We could easily determine that the key had been changed by watching the meter. This was done with idle circuits operating under traffic flow security (meaning the line was active, data equal to a constant mark, the encrypted constant mark showing on the data stream). The distribution of average voltage values (MIL STD 188) and how fast and furious it would change, hop, skip and jump were generally distinct between successive keys. DES S Box outputs have the identical symbol distribution for key and key_not (E(Rn) xor KSn, input to the S Boxes). For a given round key (and its inverse), there are between 0 and 65,536 symbols missing from the domain of the P permutation (32 bit symbols). Which symbols and how many that don't show up are dependant on the key. Some keys have no missing symbols, while others have lots. This is a function of the E permutation and R bit sharing between adjacent S Boxes. Someone appears to have been quite aware of this weakness, the second XOR operation found in a DES round ( (E(Rn) xor KSn) xor Ln )goes a long way towards masking the fact that some symbols can be missing. Were DES not to perform the second XOR, you could determine the key simply by monitoring missing symbols from the output of the S boxes (P permutation). Each new symbol found would eliminate certain patterns from the scheduled key (KSn), a 48 bit value. It would go a long way to reducing the number of unknown key bits to the range of easy brute force attacks. Now imagine that shift register based crypto generally doesn't mix key and data as well. DES operates on each bit 16 times, more than the typical shift register based crypto. Each bit of the output block of DES depends on all the input bits and all the key bits input to the key scheduler. A shift register based crypto with a shift register of a size comparable to the block size of DES would typically have a lot fewer variables contributing to each key bit, making brute force attacks on a known crypto system with known plaintext (including idle data values) much easier. Now, imagine that there is statistical significance to the output distribution of 1's and 0's based on weaker mixing. This sounds right up the alley for neural nets. Anyway, I think it really depends on the age of the crypto gear in use. Older gear tends to be less secure based on shift register size, and key/data mixing. There are also rules used to specify tap to input selections, which eliminate weak keys (the sort of rules enforced by key card readers). Attacking a cryptosystem operated with keys provided from a centralized generation/distribution system would further reduce the key search domain to strong keys. Having worked on crypto gear built the year I was born (1954) through the late '70s, I have no problem believing that Russia is using antiquated (and thus more vulnerable) crypto today. Based on replacement cost, the only thing that would drive comsec gear out of service would be demonstrable weakness (such as Bush giving Yeltsin intercepts, personnel insecurity with respect to key handling, etc.), or prohibitive maintenance/operating costs. After all, some protection is better than none.

1 0

re: real time surveillances
by Mark Hittinger 17 Dec '03

17 Dec '03

Some thoughts on this: Will the people *believe* the electronic position data? Nope. When you are busted not only will you be frisked but you will also be bulk erased :-) (to get TC's DAT tape!) The technology to fabricate video evidence is there is it not? There was a relatively recent sad case of two officers that had a video camera installed in their patrol car. They had pulled over a speeding car. While obtaining the license the occupants of the car murdered the policemen and the *entire* scene was obtained on video tape. When the tape was later reviewed, searching for a clue as to what happened, the individual had to watch the officers slowly die - all captured on video. It is possible that they might have been saved had this been real-time video rather than taped. Consider the next level here folks? Will just having archived knowledge of where certain people are be adequate? Will there not ultimately be a call for real time monitoring of these officers? Will that be something that would be cool with them? Another great idea would be anonymous real-time monitoring of vital signs along with position data. Auto-911 if you will. We know somebody at this corner had a heart attack 5 minutes ago where is he? Cheers mark.h

1 0