excerpted from: B O A R D W A T C H M A G A Z I N E Guide to the World of Online Services Editor: Jack Rickard Volume VII: Issue 11 ISSN:1054-2760 November 1993 ============== LEGALLY ONLINE ============== by Lance Rose Phil Zimmermann is a computer programmer and good citizen with an earnest desire to help others. For much of his adult life he expressed his concern through political activism, from marches in the 70's through teaching classes in military policy in the 80's. Along the way, Zimmermann became a computer network user. As the 90's began, he saw a new threat to our civil rights surface: a movement to curtail the ability of network users to discuss matters in private. Many others on the net saw the same thing, but Zimmermann actually did something about it. He melded his humanitarian ideals with his programming and cryptographic skills to create an e-mail encryption program for Everyman called Pretty Good Privacy, or PGP. He did this at no small cost to himself. An independent computer consultant, Zimmermann shelved most of his paying gigs while finishing up PGP, missing five mortgage payments on his family's home in the process. Zimmermann finished PGP in early 1991, released it in the U.S. on the Internet from his home in Colorado, and it instantly achieved worldwide popularity among network cognoscenti. PGP is free, and its customers get a lot more than they bargained for. It uses a powerful RSA public key encryption system, in which each participant owns one or more private key/public key pairs. Messages encrypted by one key in the pair can only be decrypted using the other key. The great advantage of public key encryption is that, unlike older schemes, like DES, (the current U.S. standard), you can publicly circulate the key for encrypting messages sent to you by others, since only your secret private key can decrypt those messages. When someone sends you a private message using PGP, first it encrypts the message text using the IDEA algorithm, (the International Data Encryption Algorithm, developed in Europe and comparable to DES), and generates a unique, confidential key just for decrypting the message (we'll call it the "text key"). Then he or she uses the RSA public key they received from you to encrypt the text key, and sends you both the IDEA-encrypted message and the RSA-encrypted text key. You would then use your RSA private key to decrypt the text key, and finally use the text key in turn to decrypt IDEA-encrypted message. It would be conceptually simpler to use only the RSA public and private keys for encrypting and decrypting the text, and cut out the IDEA encryption step, but there is a practical cost. Text encryption using RSA requires far more computer processing time and power than IDEA-based encryption. Zimmermann wanted to give people the full benefit of PGP without wasting time watching their computers crunch numbers, so he used the far easier to process, (but still highly secure), IDEA technique for message texts, and saved the public key RSA process for encrypting only the IDEA text keys, which are a mere 128 bits long. Thus, PGP creates securely encrypted texts, gives people the ability to distribute their public keys far and wide with no compromise in security, and extends strong encryption capabilities to those with small computers. After its release PGP became something of a community development project, and it is now in version 2.3. Other programmers and cryptographers work on enhancing the source code, creating foreign language translations of the user interface, and porting PGP from its original DOS platform to other popular computer platforms such as Mac, Vax, and most flavors of Unix, further increasing its usefulness to network users of all stripes. In the past two months, however, some dramatic new chapters swiftly opened in the ongoing saga of Phil Zimmermann and PGP. First, on the positive side, some patent-related questions nagging PGP from the very beginning finally may be close to resolution. As soon as PGP appeared, a West Coast company called Public Key Partners, headed by Jim Bidzos, claimed it violated a patent they held in computer implementations of the RSA encryption algorithm. While Bidzos did not rush into court, he did seek to suppress PGP's distribution. Among other things, he sought out the major online distribution points for PGP, such as large online services like CompuServe and GEnie BBS's including The Well, and requested that they immediately remove the PGP files from distribution because they infringed on his patents. Most services discontinued providing PGP,and it soon became an underground classic, difficult to find unless you asked the right people. Fortunately, there were many such spread across the net. Rather than challenge the patent claim, which some net observers think a worthwhile effort, Zimmermann made many requests to Bidzos to obtain a license to use RSA without hassle in PGP. Eventually, network community and industry leaders tried to obtain some sort of compromise between Zimmermann and Bidzos. Bidzos refused all entreaties and continues to oppose PGP. In the meantime, a software company named ViaCrypt obtained a license a year or two ago from Public Key Partners, (who appear willing to license virtually anyone but Zimmermann), to use the RSA algorithm in software. ViaCrypt took some time after securing the rights to figure out how it would use them. Finally, last August, it approached Zimmermann with a proposal to create a commercial version of PGP. This was a great business opportunity. ViaCrypt could use its license to legitimize PGP for the commercial market, and both could profit from PGP's high profile among companies interested in encryption. To bring PGP within ViaCrypt's license from Bidzos, Zimmermann and ViaCrypt have been replacing PGP's existing RSA encryption subroutines with comparable licensed subroutines developed by ViaCrypt. Bidzos, through his attorney, publicly expressed some doubt about whether the new ViaCrypt product will fall within its RSA license rights. Anticipating this possibility, ViaCrypt shrewdly trumped it in advance by securing a legal opinion from Brown and Bain, considered by many computer lawyers, (this author included), to be the leading computer law firm in the country, that the hybrid product is within the scope of ViaCrypt's license from Public Key Partners. The new program, tentatively brand named ViaCrypt PGP, is scheduled for rollout on November 8th of this year. If ViaCrypt PGP succeeds in the market, Zimmermann will make some money, though he was never really was in it for the money. He is also working on a new approach to the free version of PGP that may end the patent threats that continually hinder its open distribution on the net. His current efforts center on a set of encryption subroutines called RSAREF, released by Bidzos through another of his companies, RSA Data Securities Inc. (RSADSI). There is a license to the public to use the RSAREF subroutines free for noncommercial purposes: you can't make money selling it, and you can't use it for commercial messaging. By replacing PGP's custom RSA subroutines with publicly licensed equivalents from RSAREF, Zimmermann could end the patent infringement problem. RSAREF was designed to let programmers develop privacy-enhanced mail (PEM) programs using a scheme similar to that used by PGP. The text is encrypted using an easy-to-process encryption algorithm involving a single key, ending the patent infringement problem may not have been as easy at it first seemed. The text key is encrypted using the processing-intensive RSA public key scheme. The problem is that for the text encryption stage, RSADSI chose the DES algorithm instead of the IDEA algorithm used in PGP. DES was not perceived as a problem cipher until August of this year, at a cryptographers' trade show called Crypto 93. A respected Bell Northern Research scientist named Michael Weiner dropped a bombshell on the conference by asserting, in essence, that DES was dead, as a dependably secure encryption algorithm. Weiner had designed a high-speed "inside-out" DES processor chip that could test 50 million keys per second, and serve as a the basis for a highly effective DES-cracking machine. He had also priced production of the chip with a chip fabricator, and in large but not enormous quantities it would cost about $10.50 per chip. Using these figures, he said a computer with 7,000 such chips would cost a vast amount, and could find the key to any DES-encrypted message within 7 hours by testing every possible key within that time, with a mean key-cracking time of 3.5 hours per encrypted message. For $100 million, well within intelligence agency budgets, a computer could be built that would crack the keys for DES-encrypted messages at a clip of two minutes per key. This result would be achieved for texts encrypted with 56 bit DES keys, where the decryptor has a little bit of plain text he knows would be in the encrypted message, such as someone's name, or a word or two. There are DES encryption schemes using longer keys, but the 56 bit key is the U.S. government standard, and the official or de facto standard in many industry applications as well. Weiner's brute force approach would be marvelously effective despite its lack of elegance. For PEM computer programs using RSAREF in its current form, there can no longer be dependable privacy. All texts encrypted using RSAREF's standard 56 bit DES approach will, from this point forward, be vulnerable to cracking at some point by a Weiner-type supercomputer. It will be pointless to hide the DES key inside RSA encryption. Owners of the supercomputer could find the key directly from the text, and need not bother with the encrypted key. As Phil puts it, the RSA encryption of the DES key is like those little secure boxes for holding front-door keys that realtors mount on houses being shown to prospective buyers. The little box may be nearly impossible to break into without the proper code, but there's always another possibility for getting into the house without permission: break down the door or smash in the window. There's a way to avoid this with RSAREF, by invoking deeper subroutines in the package to create PEM programs that use text encryption schemes more dependable than the suddenly-reduced DES, such as the IDEA algorithm used today in PGP. IDEA uses a 128 bit key instead of the 56 bit key standard in DES, and its security has not been seriously questioned to date, despite spirited attacks by some of the world's mightiest cryptographers. Unfortunately, the current RSAREF public license only permits programmers to use the high-level routines that require DES, and prohibits using the deeper routines to bring in other encryption algorithms - the very use necessary for PGP to remain dependably secure. However, after the Weiner revelation, RSAREF will not be in much demand if it continues to restrict PEM programmers to 56 bit DES. Accordingly, it is rumored that a public license to the deeper, roll-your-own algorithm subroutines in RSAREF may soon be forthcoming from RSADSI. If this new license is issued, Zimmermann may finally be home free in his quest to create a free, effective PGP with no specter of patent infringement hovering over it. It is possible that by the end of this year or early in the next, we will see both a commercial ViaCrypt PGP, and a free PGP for personal noncommercial use. That's the good news. The bad news is that Phil Zimmermann is now the target of an investigation by the U.S. attorney's office into violations of the International Traffic in Arms Regulations, or ITAR. ITAR is a set of laws administered by the State Department, designed to keep war-grade weapons from being exported out of the U.S. to certain foreign countries. While ITAR mainly regulates weapons-like parts for tanks, jets and submarines, it also regulates encryption devices, including encryption software, as "munitions" due to their military intelligence value. The U.S. attorney is not commenting at this early stage, but observers agree the investigation relates to PGP's worldwide distribution through the Internet. This distribution constituted a clean end run around the State Department's normal procedure of placing a roadblock against all cryptography exports, until they are reviewed by the National Security Agency for military potential. What is unclear, is whether Zimmermann did anything wrong by placing PGP on the Internet on computers located within the U.S. There are very good reasons for saying that Zimmermann's actions were totally legal, and that he should not have this cloud over his head. For one thing, Zimmermann did not intend for PGP to be exported. He was never motivated to put out an international encryption standard. To the contrary, his motivation was the perception that political forces within the U.S. seemed to be pushing towards outlawing private encryption in this country. In fact, he acted specifically to put PGP into circulation quickly while it was still legal in this country, before any laws might go into place prohibiting domestic use of privately developed encryption software. For another, Internet users outside the U.S. helped themselves to PGP. Zimmermann did not send PGP anywhere outside the country. He made it available on computers within U.S. borders, which is perfectly legal in itself. By analogy, I could legally go door-to- door in this country selling devices enabling people to encrypt their telephone discussions. I can even leave an open box of them in front of my house in New Jersey, and tell all my neighbors to pick one up for their own use. If some foreign tourists take a few and spirit them back into their own countries, why should I be held guilty for export violations? In addition, Zimmermann and all the users of PGP in this country have their First Amendment rights. Zimmermann has the right to freely publish the PGP program in this country. The Constitution says Congress will enact "no law" restricting freedom of speech and of the press. There are no Amendments to the Constitution that contain exceptions for speech or press distributed through the Internet. Additionally, people who send electronic messages to each other have the right to send them encrypted without government interference, and legal action against PGP would certainly interfere with such activities. In this instance, PGP's free speech rights derive from its assistance of PGP users in exercising their own rights of free speech. This kind of derivative free speech protection is very powerful. It is analogous to the protection of speech distributors applied in the past by the Supreme Court and other federal courts to book sellers, magazine distributors, and even CompuServe. Finally, there are privacy considerations. This is not really a legal argument, as much as a question of the limits of appropriate government intrusion into peoples' private lives. The question comes up almost daily these days, in settings ranging from the privacy of employee e-mail to the swelling commercial market for extensive data on each citizen in our country of consumers. A government push against the availability of PGP, regardless of the legal cause, would count as yet another blow against the dwindling ability of us all to retain a modicum of personal privacy. As the cypherpunks are often heard to say, "if privacy becomes outlawed, only outlaws will have privacy." Powerful as these and other arguments are, they will not deter government action on their own. The government can offer the fairly standard legal argument that Zimmermann "knew or should have known," that placing PGP on the Internet would result in worldwide export in violation of the ITAR. After speaking with Zimmermann, I am not sure he actually knew, in particular, that there was a law called ITAR, or that it applied to encryption software. As mentioned above, he certainly was not out to distribute PGP worldwide. Whether the government proceeds will depend as much on political factors as on its view of the legalities involved. The investigation is at an early stage, and in fact, has not been directed formally at Zimmermann. The only activity in public view so far was the service of subpoenas for document production by the U.S. Attorney's office in San Jose, CA on Viacrypt in Phoenix, AZ and another company named Austin Code Works in Austin, TX. Austin Code Works distributes PGP and other free software for encryption and other uses in source code form, for little more than the price of a computer disk. According to Zimmermann, he has no business relationship with Austin Code Works. He also had no idea they were distributing PGP until he read about the subpoena served on them. As soon as Austin Code Works was served, its president, Grady Ward, went public on the Internet with a ringing defense of its position. Ward claims they do not distribute executable programs, but only "source code algorithmic descriptions" of encryption techniques, thus falling under a "technical data" exception to ITAR. The State Department publicly countered that position, and is requiring Austin Code Works to register as a munitions dealer. There is no telling whether the investigation will proceed to charges against Zimmermann or others, but Zimmermann and others in the network community intend to be prepared. Phil Zimmermann's attorney, Colorado criminal lawyer Philip Dubois, is accepting contributions for Zimmermann's defense, (He can be reached at Philip Dubois, Esq., 2305 Broadway, Boulder, CO 80304, (303)444-3885, dubois@csn.org). The Electronic Frontier Foundation is also stepping forward in Zimmermann's defense, with financial commitments from EFF and several of its individual board members, and efforts to rally public support for Zimmermann and PGP. A lingering question in the current investigation is why the government waited over two years after PGP's release to start it up. Some speculate it is due to a link between the investigation and the government's efforts to establish a new encryption standard named CLIPPER as the replacement for the aging DES standard. The government has repeatedly stated it will not seek to make Clipper the only legal encryption standard in this country by outlawing all others. But if it proceeds to charge Zimmermann and PGP as a result of the current investigation, it could have the effect of using the government's legal artillery to blow away one of Clipper's most prominent competitors. Speculation aside, PGP's legal situation is slowly maturing, and within another year or so we should know for sure whether it's legal or illegal in the U.S. By that time, it will be in the hands of millions of people the world over, each using PGP to create his or her own private communications channel. Hopefully, we will not have to witness the ironic spectacle of PGP being banned in the country of its birth, while freely in use in the rest of the world. [Lance Rose is an attorney practicing high-tech and information law in Montclair, NJ. He can be found on the Internetat elrose@well.sf.ca.us, and on Compuserve at 72230,2044. He is also author of SysLaw, the legal guide for online service providers, available from PC Information Group at 800-321-8285. Pretty Good Privacy is available in it's latest July 1, 1993 release as PGP23A.ZIP with C language source code in PGP23ASR.ZIP. Phil Zimmermann can be reached at Boulder Software Engineering, 3021 Eleventh Street, Boulder, Colorado 80304; (303)541-0140 voice/fax; or via Internet at prz@acm.org ViaCrypt will make the commercial PGP available at an intro price of $100. ViaCrypt, 2104 W. Peoria, Phoenix, AZ 80209, (602) 944-1543. - Editor