- cypherpunks-legacy - lists.cpunks.org

Re: Commercial PGP: Verifying Trustworthiness
by doug＠netcom5.netcom.com 31 Aug '93

31 Aug '93

--- Forwarded mail from Marc Horowitz <marc(a)Athena.MIT.EDU> >From owner-cypherpunks(a)toad.com Mon Aug 30 23:40:01 1993 Received: from relay2.UU.NET by mail.netcom.com (5.65/SMI-4.1/Netcom) id AA14421; Mon, 30 Aug 93 23:39:57 -0700 Received: from toad.com by relay2.UU.NET with SMTP (5.61/UUNET-internet-primary) id AA10745; Tue, 31 Aug 93 02:40:50 -0400 Received: by toad.com id AA14781; Mon, 30 Aug 93 23:33:56 PDT Received: by toad.com id AA14701; Mon, 30 Aug 93 23:31:26 PDT Return-Path: <marc(a)Athena.MIT.EDU> Received: from Athena.MIT.EDU ([18.72.1.1]) by toad.com id AA14688; Mon, 30 Aug 93 23:31:23 PDT Received: from OLIVER.MIT.EDU by Athena.MIT.EDU with SMTP id AA00837; Tue, 31 Aug 93 02:28:59 EDT Received: by oliver.MIT.EDU (AIX 3.2/UCB 5.64/4.7) id AA14903; Tue, 31 Aug 1993 02:28:52 -0400 Message-Id: <9308310628.AA14903(a)oliver.MIT.EDU> To: bbyer(a)BIX.com Cc: honey(a)citi.umich.edu, cypherpunks(a)toad.com Subject: Re: Commercial PGP: Verifying Trustworthiness In-Reply-To: Your message of Tue, 31 Aug 93 00:14:18 -0400. <9308310014.memo.72462(a)BIX.com> Date: Tue, 31 Aug 93 02:28:52 EDT From: Marc Horowitz <marc(a)Athena.MIT.EDU> Marc Horowitz <marc(a)Athena.MIT.EDU> said: >> I dunno. The early versions of UNIX had a back door in the login [...] >I've let a lot of stupid comments go by, but I have to respond to this one. > >It is true that Dennis Ritchie (I believe, if not him, one of the >other original UNIX authors) proposed such a login/compiler virus. >But it wasn't in any early version of UNIX. Stupid? Watch the flame bait...he merely overstated a touch. The back doors weren't part of any of the full distributions, it's true, but they were quite a bit more than proposals. Ken Thompson actually distributed those back doors via a compiler update, warning of a security problem and urging all sites to recompile. Most did, which inserted the back doors into the programs. That's close enough to the original claim. See the Ken Thompson & Dennis Ritchie Turing Award Lecture, which goes into detail about this. The level of sneakiness involved was amazing. Compilers are the ultimate security breach. Doug

1 0

Re: Apple, privacy, and AOCE
by collins＠newton.apple.com 31 Aug '93

31 Aug '93

>I found the idea that RSADSI will be generating folks' key pairs >particularly chilling. What I gathered from actually using this software is that you personally generate a key pair, on your own machine, and then transparently send your public key to RSADSI. Some time later, you receive a certificate (with an expiration date) that allows your 'signer' to function. RSADSI does not make, or even see, your private key. >The article accompanying these sidebars suggests that folks' private keys >will be stored on the server; My understanding is that address books on the [optional] servers may have copies of certificates, for people who have certificates and want them published. >the article [...] must be the result of miscommunication yes >The NSA recently signed an agreement with the Software Publishers >Association that will provide expedited approval of RC-4 encryption based >on 40-bit keys. Not surprising, since a pre-computation attack allowing a direct key lookup against RC-4 with 40 bit keys is economically feasible for anyone who can afford a CD-ROM jukebox (128 mips-years of computation + 8 terabytes of storage). >NSA [...] will allow slightly more-powerful scrambling capabilities >[in AOCE] AOCE uses 64-bit keys, and larger keys mean better security. This could mean anything. They might actually be using 64 bit keys (which would be good, although 80 bits is recommended), or they might be using 40 bit keys with 24 bits of salt (or worse: 32 and 32). Salted keys (key+salt of sufficient size), stop the precomputed attack, but if the actual key size, without salt, is still only 40 bits, then exhaustive search of the keyspace, after the salt has been seen, will only take 64 mips-years. Scott Collins | "Few people realize what tremendous power there | is in one of these things." -- Willy Wonka ......................|................................................ BUSINESS. voice:408.862.0540 fax:974.6094 collins(a)newton.apple.com Apple Computer, Inc. 1 Infinite Loop, MS 301-2C Cupertino, CA 95014 ....................................................................... PERSONAL. voice/fax:408.257.1746 1024:669687 catalyst(a)netcom.com

1 0

Internet commerce...for your cracking pleasure
by nobody＠shell.portal.com 31 Aug '93

31 Aug '93

================================================= INDIVIDUAL INTERNET SUBSCRIPTIONS AVAILABLE! Have you ever wanted the convenience of the news delivered directly to your e-mail box? Now its possible! ...USA Today....The Moscow News is available for just $80 a year. For more information, or to place an order send an e-mail to subscrib(a)americast.com Include the e-mail the address to which you want the services delivered. In that e-mail please designate the paper and sections you would like to receive, give your full name, address, phone number, and the credit card (Visa or MasterCard) you will be using for the purchase, and the credit card expiration date.

1 0

Re: Commercial PGP: Verifying Trustworthiness
by bbyer＠BIX.com 31 Aug '93

31 Aug '93

In-Reply-To: <9308272026.AA17010(a)toad.com> > From: peter honeyman <honey(a)citi.umich.edu> > trust? you could read the code, starting at about line 550 of crypto.c. > of course, you have to trust your eyes, your editor (if you use one), > and your operating system not to deceive you. (i think i've carried > this too far.) I dunno. The early versions of UNIX had a back door in the login program put in by the designer. The compiler also watched for the login source code to be recompiled and added the back door. The compiler also watched for the compiler source code to be recompiled and inserted the login code modification code _and the compiler modification code. You can never be to careful.

2 1

Diffie Hellman
by Timothy Newsham 31 Aug '93

31 Aug '93

Hi, Does anyone know of DH code that is freely available and distributable? Preferably available outside the USA. Preferably speed optimized. although I'll put up w/ whatever i can get my hands on. Tim N.

1 0

Apple, privacy, and AOCE
by greg＠ideath.goldenbear.com 31 Aug '93

31 Aug '93

Recent discussion about Apple's commitment (or lack thereof) to providing anything like real privacy calls to mind some stuff from MacWeek of several weeks ago; the 7/12/93 issue of MacWeek includes a "special report" on AOCE (Apple Open Collaborative Environment, a "groupware" (anyone have any better explanation?) setup which facilitates sharing data between colleagues. The "special report" includes two sidebars about encryption and security in the AOCE environment, which I reproduce below. I called and asked about permission to reprint the entire article to the list but was unable to get past marketroids who wanted me to pay $50 + copying costs (payable to them) for my 4-color glossy reprints showcasing my product. (don't have one, dammit.) Some folks still don't "get" the net. Sigh. I found the idea that RSADSI will be generating folks' key pairs particularly chilling. The article accompanying these sidebars suggests that folks' private keys will be stored on the server; the article made the security of the thing sound *so* poorly designed that I figure it must be the result of miscommunication between the Apple folks and the article's author (Mitch Radcliffe). If anyone really cares I can see about posting some more of the article. --- "Apple could have the toughest code on the block" The Cold War mind-set in Washington, D.C., thawed a bit when Apple gained permission from the U.S. government to ship AOCE outside the United States while retaining the software's advanced encryption features. Encryption technology is considered munitions by the national defense apparatus. For many years, the Department of Commerce, in collaboration with the National Security Agency, has limited encryption technology exports. But the strict limits on cryptography seem to be eroding after the fall of the Soviet Union. According to sources, the RC-4 encryption in AOCE was approved under a special agreement between Apple and the NSA that will allow slightly more-powerful scrambling capabilities than are typically given export clearance. The NSA recently signed an agreement with the Software Publishers Association that will provide expedited approval of RC-4 encryption based on 40-bit keys. AOCE uses 64-bit keys, and larger keys mean better security. "Protection can be at many levels," Gursharan Sidhu, Apple technical director of collaboration systems development, told attendees at the Apple Worldwide Developers Conference in May. He said it's very easy to protect against casual intrusion and even determined hackers. But a resource-rich intruder with access to supercomputers can defeat many encryption technologies. "You've got to look at these various dimensions of how far you want to go in providing security," Sidhu said. Apple's security is among the best available, he said. The company fought for several years to win approval for AOCE's privacy and security technologies. "In the world of commercial security, it's a matter of saying it is secure within the parameters of commercial reality," Sidhu said. Sources said it would take a supercomputer-equipped intruder from two hours to six days to crack a 64-bit RC-4 key using a brute-force attack that analyzed the encrypted data for hints about the key. Even if such an intruder cracked the key for one network session, each successive session would have to be recracked because it would have a different key. Apple will be constrained from selling AOCE in countries on the State Department's list of terrorist nations. And the company will sell a version of AOCE without encryption functions in France, because the French government requires access to all imported encryption technology. In the United States, the barriers against encryption export seem to be falling. In May, the Computer System Security and Privacy Advisory Board of the National Institute of Standards and Technology, the Department of Commerce agency that monitors civilian encryption technologies, issued resolutions that recommend the United States revise its export laws to facilitate the spread of encryption technologies. "I think the NIST resolutions are a good indication that the Department of Commerce is trying to make it clear to the White House that these policies are backward," said Marc Rotenberg, director of the Washington, D.C., office of Computer Professionals for Social Responsibility. "If they don't change, they'll still be making policy for the 1950's." - by Mitch Ratcliffe --------- "Behind it all, digital signatures" If the messaging engine is the workhorse of AOCE, the digital signature capability is the electronic Paul Revere that carries the important information. A digital signature is an electronic analogy to the written signature, and no two are alike. The result of complex cryptographic processes, a digital signature can prove that a particular user "signed" a particular document. Users sign a document by clicking a check box in the AOCE mailer. The legal force of such signatures has not been determined, but Apple believes they are sufficiently reliable for building audit trails through a company or even between companies. Users will have to apply for and recieve their digital signatures from RSA Data Security Inc. of Redwood City, Calif., which developed the technology. The signature arrives in two parts, a private key that the user must keep secure and a public key that can be distributed freely. Any document signed using the private key can be compared with the user's public key to see if the document is authentic and unchanged. Public keys can be handed out by a user or stored in a key database, where users can go to get keys that have been verified by a third party. For example, a bank might keep the list of public keys that can be used by a customer's company for signing purchase orders. Developers, such as Snow Development Corp. of Clearwater, Fla., and Shana Corp. of Edmonton, Alberta, plan to use AOCE digital signatures in report-routing and forms software. The applications will let a user create a flowchart of colleagues who need to sign off on a document, send the document to that list serially or in parallel, and collect all the digital signatures for auditing purposes in the final document. -- by Mitch Ratcliffe -- Greg Broiles greg(a)goldenbear.com Golden Bear Computer Consulting +1 503 342 7982 Box 12005 Eugene OR 97440 BBS: +1 503 687 7764

1 0

Re: Apple+Clipper
by collins＠newton.apple.com 31 Aug '93

31 Aug '93

Tim and Paul present accurate evidence mitigating some of my specific points. I agree that the AV Macs and AOCE are steps (or even leaps) towards a platform favorable for clipper infestation. .....BEGIN HAIR SPLITTING MESSAGE..... I would like to fine tune one of Tim's comments: >#1 and #2 are already satisfied with the new generation of av Macs, as These machines most certainly do not contain the clipper chip (which I'm sure Tim did not mean to imply), though they do show that Apple can make machines with special purpose hardware and capabilities not present, or emulatable, on earlier machines (this, I think, was Tim's point: a counter example to my speculation). It is illuminating to note, however, that the 880av is not based on the same hardware as the 800. It is an earlier generation that took longer to come to market. In this sense it recalls the late IIfx i.e., faster for a while and then a dead end. In many ways, the 800 is a superior machine, and the 880 has some catching up to do (like adding interleaved RAM access, for one). By my count, the 880 has taken over three years to come to market. Summary: the AV Macs may indicate a new trend, but they do not represent an immediate foothold for clipper. .....END HAIR SPLITTING MESSAGE..... Other than that, and although they come to slightly different conclusions, I agree with Tim and Paul, who examined the root motives rather than 'diagnosing for symptoms'. My final take on this is: Apples history and our guesses about its likely motives lead me to (still) predict that Apple can't immediately jump on the Clipper bandwagon; though the further out we speculate (beyond 2 years?), the less faith in this prediction we can justify. I like to be wrong almost as much as I like to be right. More, if the pay is better. Scott Collins | "Few people realize what tremendous power there | is in one of these things." -- Willy Wonka ......................|................................................ BUSINESS. voice:408.862.0540 fax:974.6094 collins(a)newton.apple.com Apple Computer, Inc. 1 Infinite Loop, MS 301-2C Cupertino, CA 95014 ....................................................................... PERSONAL. voice/fax:408.257.1746 1024:669687 catalyst(a)netcom.com

2 2

Help/Advice requested
by Forrest Aldrich 31 Aug '93

31 Aug '93

I'm working on the development of a resonably secure software registration scheme, and thought that someone from the Cypherpunk list would be ideal for helping out. Sample code is even better, but general advice to do with the methodology and coding would be most appreciated. There is a program out there called PROTECT! EXE/COM, which does something on the lines of EXE encryption which has been brought up here. I've corresponded with the author and he won't release the code (understandably, he's worked very hard on it). Apparently what it does is (after you have run something like PKLITE on your executable) encapsulates the EXE file, encrypted, with CRC checks and debugger traps. He claims this is quite effective and has not had anyone come through with cracking it yet. Just thought I would mention this .... Unfortunately, that program is only for DOS/Windows. It would be NICE to have something like this working under UNIX, and I can't imagine that it would be that difficult, although it would most certainly be platform (assembly language) specific. Anyone have some ideas about how to implement such a beast? Thanks alot...

1 0

AV Macs and Apple Paranoia
by Jamie Jamison 31 Aug '93

31 Aug '93

Actually I was going to ask in here if the DSP in the new Macs would lend itself to real-time PGP use for voice communications. WITH STANDARD DISCLAIMER

2 1

Re: Matrix extensions of the rsa algorithm. (fwd)
by Timothy Newsham 30 Aug '93

30 Aug '93

Forwarded message: > You've probably gotten some other responses about this... > > I think that the paper you describe has been implemented as the Warlock > cyphersystem. Someone else posted a C++ Warlock implementation, and > some related material, to cypherpunks about two months ago. I can't > find my filed copy, but if you ask I'm sure someone has it stashed > away. > > Paul Robichaux, KD4JZG | "Change the world for a better tomorrow. But > perobich(a)ingr.com | watch your ass today." - aaron(a)halcyon.com > Intergraph Federal Systems | Be a cryptography user- ask me how. Does anyone have a copy of Warlock? I'd appreciate getting a copy from someone.

1 0