The legal landscape of cryptography is complex and constantly changing. In recent years the legal restrictions on cryptography in the United States have largely eased, while the restrictions in other countries have increased somewhat.
In this section, we’ll examine restrictions that result from patent law, trade secret law, import/export restrictions, and national security concerns.
These regulations and laws are in a constant state of change, so be sure to consult with a competent attorney (or three) if you will be using cryptography commercially or internationally.
Some of the earliest and most important software patents granted by the U.S. Patent and Trademark Office were in the field of cryptography. These software patents go back to the late 1960s and early 1970s. Although computer algorithms were widely thought to be unpatentable at the time, the cryptography patents were allowed because they were written as patents on encryption devices that were built with hardware—computers at the time were too slow to perform meaningfully strong encryption in a usably short time. IBM’s original patents on the algorithm that went on to become the U.S. Data Encryption Standard (DES) were, in fact, on a machine that implemented the encryption technique.
The doctrine of equivalence holds that if a new device operates in substantially the same way as a patented device and produces substantially the same result, then the new device infringes the original patent. As a result of this doctrine, which is one of the foundation principles of patent law, a program that implements a patented encryption technique will violate that patent, even if the original patent was on a machine built from discrete resistors, transistors, and other components. Thus, the advent of computers that were fast enough to implement basic logic circuits in software, combined with the acceptance of patent law and patents on electronic devices, assured that computer programs would also be the subject of patent law.
The underlying mathematical techniques that cover nearly all of the cryptography used on the Web today are called public key cryptography . This mathematics was largely worked out during the 1970s at Stanford University and the Massachusetts Institute of Technology. Both universities applied for patents on the algorithms and for the following two decades the existence of these patents was a major barrier to most individuals and corporations that wished to use the technology.
Today the public key patents have all expired. For historical interest, here they are:
Martin E. Hellman and Ralph C. Merkle
Expired August 19, 1997
The Hellman-Merkle patent introduced the underlying technique of public key cryptography and described an implementation of that technology called the knapsack algorithm . Although the knapsack algorithm was later found to be not secure, the patent itself withstood several court challenges throughout the 1990s.
Martin E. Hellman, Bailey W. Diffie, and Ralph C. Merkle
Expired April 29, 1997
 Ronald L. Rivest, Adi Shamir, and Leonard M. Adleman
Expired September 20, 2000.
Many other algorithms beyond initial public key algorithms have been patented. For example, the IDEA encryption algorithm used by PGP Version 2.0 is covered by U.S. patent 5,214,703; its use in PGP was by special arrangement with the patent holders, the Swiss Federal Institute of Technology (ETH) and a Swiss company called Ascom-Tech AG. Because this patent does not expire until May 2011, the algorithm has not been widely adopted beyond PGP.
In addition to algorithms, cryptographic protocols can be protected by patent. In general, one of the results of patent protection appears to be reduced use of the invention. For example, deployment of the DigiCash digital cash system has been severely hampered by the existence of several patents (4,529,870, 4,759,063, 4,759,064, and others) that prohibit unlicensed implementations or uses of the technology.
Although new encryption algorithms that are protected by patents will continue to be invented, a wide variety of unpatented encryption algorithms now exist that are secure, fast, and widely accepted. Furthermore, there is no cryptographic operation in the field of Internet commerce that requires the use of a patented algorithm. As a result, it appears that the overwhelming influence of patent law in the fields of cryptography and e-commerce have finally come to an end.
Until very recently, many nontechnical business leaders mistakenly believed that they could achieve additional security for their encrypted data by keeping the encryption algorithms themselves secret. Many companies boasted that their products featured proprietary encryption algorithms. These companies refused to publish their algorithms, saying that publication would weaken the security enjoyed by their users.
Today, most security professionals agree that this rationale for keeping an encryption algorithm secret is largely incorrect. That is, keeping an encryption algorithm secret does not significantly improve the security that the algorithm affords. Indeed, in many cases, secrecy actually decreases the overall security of an encryption algorithm.
There is a growing trend toward academic discourse on the topic of cryptographic algorithms. Significant algorithms that are published are routinely studied, analyzed, and occasionally found to be lacking. As a result of this process, many algorithms that were once trusted have been shown to have flaws. At the same time, a few algorithms have survived the rigorous process of academic analysis. These are the algorithms that are now widely used to protect data.
Some companies think they can short-circuit this review process by keeping their algorithms secret. Other companies have used algorithms that were secret but widely licensed as an attempt to gain market share and control.
But experience has shown that it is nearly impossible to keep the details of a successful encryption algorithm secret. If the algorithm is widely used, then it will ultimately be distributed in a form that can be analyzed and reverse-engineered.
Consider these examples:
In the 1980s Ronald Rivest developed the RC2 and RC4 data encryption algorithms as an alternative to the DES. The big advantage of these algorithms was that they had a variable key length—the algorithms could support very short and unsecure keys, or keys that were very long and thus impossible to guess. The United States adopted an expedited export review process for products containing these algorithms, and as a result many companies wished to implement them. But RSA kept the RC2 and RC4 algorithms secret, in an effort to stimulate sales of its cryptographic toolkits. RSA Data Security widely licensed these encryption algorithms to Microsoft, Apple, Lotus, and many other companies. Then in 1994, the source code for a function claiming to implement the RC4 algorithm was anonymously published on the Internet. Although RSA Data Security at first denied that the function was in fact RC4, subsequent analysis by experts proved that it was 100 percent compatible with RC4. Privately, individuals close to RSA Data Security said that the function had apparently been “leaked” by an engineer at one of the many firms in Silicon Valley that had licensed the source code.
Likewise, in 1996 the source code for a function claiming to implement the RC2 algorithm was published anonymously on the Internet. This source code appeared to be derived from a copy of Lotus Notes that had been disassembled.
In the 1990s, the motion picture industry developed an encryption system called the Contents Scrambling System (CSS) to protect motion pictures that were distributed on DVD (an acronym that originally meant Digital Video Discs but now stands for Digital Versatile Discs in recognition of the fact that DVDs can hold information other than video images). Although the system was thought to be secure, in October 1999 a C program that could descramble DVDs was posted anonymously to an Internet mailing list. The original crack was due to a leaked key. Since then, cryptographers have analyzed the algorithm and found deep flaws that allow DVDs to be very easily descrambled. Numerous programs that can decrypt DVDs have now been published electronically, in print, on posters, and even on t-shirts.
Because such attacks on successful algorithms are predictable, any algorithm that derives its security from secrecy is not likely to remain secure for very long. Nevertheless, algorithms that are not publicly revealed are protected by the trade secret provisions of state and federal law.
In the past 50 years there has been a growing consensus among many governments of the world on the need to regulate cryptographic technology. The original motivation for regulation was military. During World War II, the ability to decipher the Nazi Enigma machine gave the Allied forces a tremendous advantage—Sir Harry Hinsley estimated that the Allied “ULTRA” project shortened the war in the Atlantic, Mediterranean, and Europe “by not less than two years and probably by four years.” As a result of this experience, military intelligence officials in the United States and the United Kingdom decided that they needed to control the spread of strong encryption technology—lest these countries find themselves in a future war in which they could not eavesdrop on the enemy’s communications.
In the early 1990s, law enforcement organizations began to share the decades’ old concerns of military intelligence planners. As cryptographic technology grew cheaper and more widespread, officials worried that their ability to conduct search warrants and wiretaps would soon be jeopardized. Agents imagined seizing computers and being unable to read the disks because of encryption. They envisioned obtaining a court-ordered wiretap, attaching their (virtual) alligator clips to the suspect’s telephone line, and then hearing only static instead of criminal conspiracies. Their fears were made worse when crooks in New York City started using cellular phones, and the FBI discovered that the manufacturers of the cellular telephone equipment had made no provisions for executing wiretaps. Experiences like these strengthened the resolve of law enforcement and intelligence officials to delay or prevent the widespread adoption of unbreakable cryptography.
Export controls in the United States are enforced through the Defense Trade Regulations (formerly known as the International Traffic in Arms Regulation—ITAR). In the 1980s, any company wishing to export a machine or program that included cryptography needed a license from the U.S. government. Obtaining a license could be a difficult, costly, and time-consuming process.
In 1992, the Software Publishers Association opened negotiations with the State Department to create an expedited review process for mass-market consumer software. The agreement reached allowed the export of programs containing RSA Data Security’s RC2 and RC4 algorithms, but only when the key size was set to 40 bits or less.
Nobody was very happy with the 40-bit compromise. Many commentators noted that 40 bits was not sufficient to guarantee security. Indeed, the 40-bit limit was specifically chosen so the U.S. government could decrypt documents and messages encrypted with approved software. Law enforcement officials, meanwhile, were concerned that even 40-bit encryption could represent a serious challenge to wiretaps and other intelligence activities if the use of the encryption technology became widespread.
Following the 1992 compromise, the Clinton Administration made a series of proposals designed to allow consumers the ability to use full-strength cryptography to secure their communications and stored data, while still providing government officials relatively easy access to the plaintext, or the unencrypted data. These proposals were all based on a technique called key escrow. The first of these proposals was the administration’s Escrowed Encryption Standard (EES), more commonly known as the Clipper chip.
The idea behind key escrow is relatively simple. Whenever the Clipper chip is asked to encrypt a message, it takes a copy of the key used to encrypt that message, encrypts that message key with a second key, and sends a copy of the encrypted key with the encrypted message. The key and message are stored in a block of data called the Law Enforcement Access Field (LEAF). Each Clipper chip has a second key that is unique to that particular Clipper chip, and the government has a copy of every one of them (or the ability to recreate them). If the Federal Bureau of Investigation, the National Security Agency, or the Central Intelligence Agency should intercept a communication that is encrypted with a Clipper chip, the government simply determines which key was used to encrypt the message, gets a copy of that chip’s second key, and uses the second key to decrypt the LEAF. First, it decrypts the message key, and then uses the message key to decrypt the message. This is shown schematically in Figure 4-2.
Figure 4-2. A copy of the key used to encrypt the message is included with the encrypted message in a block of data called the Law Enforcement Access Field (LEAF). Each LEAF is encrypted with a key that is unique to the particular Clipper chip. (Adapted with permission from a diagram prepared by SEARCH.)
The Clipper chip proposal was met with much hostility from industry and the general public (see Figure 4-3). Criticism centered upon two issues. The first objection was hardware oriented: PC vendors said that it would be prohibitively expensive to equip computers with a Clipper chip. The second objection was philosophical: many people thought that the government had no right mandating a back door into civilian encryption systems; Clipper was likened to living in a society where citizens had to give the police copies of their front door keys.
In early 1996, the Clinton Administration proposed a new system called software key escrow . Under this new system, companies would be allowed to export software that used keys up to 64 bits in size, but only under the condition that a copy of the key used by every program had been filed with an appropriate escrow agent within the United States, so that if law enforcement required, any files or transmission encrypted with the system could be easily decrypted. This proposal, in turn, was replaced with another one called key recovery , that didn’t require escrowing keys at all, but that still gave the government a back door to access encrypted data. As an added incentive to adopt key recovery systems, the Clinton Administration announced that software publishers could immediately begin exporting mass-market software based on the popular DES algorithm (with 56 bits of security) if they committed to developing a system that included key recovery with a 64-bit encryption key.
Figure 4-3. The “big brother inside” campaign attacked the Clinton Administration’s Clipper chip proposal with a spoof on the successful “intel inside” marketing campaign.
None of these programs were successful. They all failed because of a combination of technical infeasibility, the lack of acceptance by the business community, and objections by both technologists and civil libertarians.
In 1997, an ad hoc group of technologists and cryptographers issued a report detailing a number of specific risks regarding all of the proposed key recovery, key escrow, and trusted third-party encryption schemes (quotations are from the report):
There is fundamentally no way to prevent the compromise of the system “by authorized individuals who abuse or misuse their positions. Users of a key recovery system must trust that the individuals designing, implementing, and running the key recovery operation are indeed trustworthy. An individual, or set of individuals, motivated by ideology, greed, or the threat of blackmail, may abuse the authority given to them. Abuse may compromise the secrets of individuals, particular corporations, or even of entire nations. There have been many examples in recent times of individuals in sensitive positions violating the trust placed in them. There is no reason to believe that key recovery systems can be managed with a higher degree of success.”
Securing a communications or data storage system is hard work; the key recovery systems proposed by the government would make the job of security significantly harder because more systems would need to be secured to provide the same level of security.
The envisioned key recovery system would have to work with thousands of products from hundreds of vendors; it would have to work with key recovery agents all over the world; it would have to accommodate tens of thousands of law enforcement agencies, tens of millions of public-private key pairs, and hundreds of billions of recoverable session keys: “The overall infrastructure needed to deploy and manage this system will be vast. Government agencies will need to certify products. Other agencies, both within the U.S. and in other countries, will need to oversee the operation and security of the highly-sensitive recovery agents—as well as ensure that law enforcement agencies get the timely and confidential access they desire. Any breakdown in security among these complex interactions will result in compromised keys and a greater potential for abuse or incorrect disclosures.”
A functioning key recovery system would deal with hundreds of requests for keys every week coming from many difficult sources. How could all these requests be properly authenticated?
Operating a key recovery system would be incredibly expensive. These costs include the cost of designing products, engineering the key recovery center itself, actual operation costs of the center, and (we hope) government oversight costs. Invariably, these costs would be passed along to the end users, who would be further saddled with “both the expense of choosing, using, and managing key recovery systems and the losses from lessened security and mistaken or fraudulent disclosures of sensitive data.”
On July 7, 1998, the U.S. Bureau of Export Administration (BXA) issued an interim rule giving banks and financial institutions permission to export non-voice encryption products without data-recovery features, after a one-time review. This was a major relaxation of export controls, and it was just a taste of what was to come. On September 16, 1998, the BXA announced a new policy that allowed the export of any program using 56-bit encryption (after a one-time review) to any country except a “terrorist country,” export to subsidiaries of U.S. companies, export to health, medical, and insurance companies, and export to online merchants. This policy was implemented by an interim rule on December 31, 1998.
In August 1999, the President’s Export Council Subcommittee on Encryption released its “Liberalization 2000” report, that recommended dramatic further relaxations of the export control laws. The Clinton Administration finally relented, and published new regulations on January 12, 2000. The new regulations allowed export of products using any key length after a technical review to any country that was not specifically listed as a “terrorist country.” The regulations also allowed unrestricted exporting of encryption source code without technical review provided that the source code was publicly available. Encryption products can be exported to foreign subsidiaries of U.S. firms without review. Furthermore, the regulations implemented the provisions of the Wassenaar Arrangement (described in Section 22.214.171.124) allowing export of all mass-market symmetric encryption software or hardware up to 64 bits. For most U.S. individuals and organizations intent on exporting encryption products, the U.S. export control regulations were essentially eliminated.
In the final analysis, the U.S. export control regulations were largely successful at slowing the international spread of cryptographic technology. But the restrictions were also successful in causing the U.S. to lose its lead in the field of cryptography. The export control regulations created an atmosphere of both suspicion and excitement that caused many non-U.S. academics and programmers to take an interest in cryptography. The regulations created a market that many foreign firms were only too happy to fill. The widespread adoption of the Internet in the 1990s accelerated this trend, as the Internet made it possible for information about cryptography—from historical facts to current research—to be rapidly disseminated and discussed. It is no accident that when the U.S. National Institute of Standards and Technology decided upon its new Advanced Encryption Standard, the algorithm chosen was developed by a pair of Europeans. The U.S. export controls on cryptography were ultimately self-defeating. Nevertheless, an ongoing effort to regulate cryptography at the national and international level remains.
The Crypto Law Survey is a rather comprehensive survey of these competing legislative approaches compiled by Bert-Jaap Koops. The survey can be found on the Web at the location http://cwis.kub.nl/~frw/people/koops/lawsurvy.htm. The later sections on international laws, as well as parts from the preceding paragraphs, draw from Koops’ January 2001 survey, which was used with his permission.
There is a huge and growing market in digital media. Pictures, e-books, music files, movies, and more represent great effort and great value. Markets for these items are projected to be in the billions of dollars per year. The Internet presents a great medium for the transmission, rental, sale, and display of this media. However, because the bits making up these items can be copied repeatedly, it is possible that fraud and theft can be committed easily by anyone with access to a copy of a digital item. This is copyright violation, and it is a serious concern for the vendors of intellectual property. Vendors of digital media are concerned that unauthorized copying will be rampant, and the failure of users to pay usage fees would be a catastrophic loss. The experience with Napster trading of MP3 music in 2000-2001 seemed to bear out this fear.
Firms creating and marketing digital media have attempted to create various protection schemes to prevent unauthorized copying of their products. By preventing copying, they would be able to collect license fees for new rentals, and thus preserve their expected revenue streams. However, to date, all of the software-only schemes in general use (most based on cryptographic methods) have been reverse-engineered and defeated. Often, the schemes are actually quite weak, and freeware programs have been posted that allow anyone to break the protection.
In response to this situation, major intellectual property creators and vendors lobbied Congress for protective legislation. The result, the Digital Millennium Copyright Act (DMCA), was passed in 1998, putatively as the implementation of treaty requirements for the World Intellectual Property Organization’s international standardization of copyright law. However, the DMCA went far beyond the requirements of the WIPO, and added criminal and civil provisions against the development, sale, trafficking, or even discussion of tools or methods of reverse-engineering or circumventing any technology used to protect copyright. As of 2001, this has been shown to have unexpected side-effects and overbroad application. For example:
A group of academic researchers were threatened with a lawsuit by the music industry for attempting to publish a refereed, scientific paper pointing out design flaws in a music watermarking scheme.
A Russian programmer giving a security talk at a conference in the U.S. was arrested by the FBI because of a program manufactured and sold by the company for which he worked. The creation and sale of the program was legal in Russia, and he was not selling the program in the U.S. while he was visiting. He faces (as of 2001) significant jail time and fines.
A publisher has been enjoined from even publishing URLs that link to pages containing software that can be used to play DVDs on Linux systems. The software is not industry-approved and is based on reverse-engineering of DVDs.
Several notable security researchers have stopped research in forensic technologies for fear that their work cannot be published, or that they might be arrested for their research—despite those results being needed by law enforcement.
Traditional fair use exemptions for library and archival copies of purchased materials cannot be exercised unless explicitly enabled in the product.
Items that pass into the public domain after the copyright expires will not be available, as intended in the law, because the protection mechanisms will still be present.
Narrow exemptions were added to the law at the last moment allowing some research in encryption that might be used in copyright protection, but these exemptions are not broad enough to cover the full range of likely research. The exemptions did not cover reverse-engineering, in particular, so the development and sale of debuggers and disassemblers could be considered actionable under the law.
The DMCA is under legal challenge in several federal court venues as of late 2001. Grounds for the challenges include the fact that the law chills and imposes prior restraint on speech (and writing), which is a violation of the First Amendment to the U.S. Constitution.
If the law stands, besides inhibiting research in forensic tools (as mentioned above), it will also, paradoxically, lead to poorer security. If researchers and hobbyists are prohibited by law from looking for weaknesses in commonly used protocols, it is entirely possible that weak protocols will be deployed without legitimate researchers being able to alert the vendors and public to the weaknesses. If these protocols are also used in safety- and security-critical applications, they could result in danger to the public: criminals and terrorists will not likely be bound by the law, and might thus be in a position to penetrate the systems.
Whether or not the DMCA is overturned, it is clear that many of the major entertainment firms and publishers will be pressing for even greater restrictions to be legislated. These include attempts to put copyright protections on arbitrary collections of data and attempts to require hardware support for copyright protection mechanisms mandated in every computer system and storage device. These will also likely have unintended and potentially dangerous consequences.
As this is a topic area that changes rapidly, you may find current news and information at the web site of the Association for Computing Machinery (ACM) U.S. Public Policy Committee (http://www.acm.org/usacm/ ). The USACM has been active in trying to educate Congress about the dangers and unfortunate consequences of legislation that attempts to outlaw technology instead of penalizing infringing behavior. However, the entertainment industry has more money for lobbying than the technical associations, so we can expect more bad laws in the coming years.
International agreements on the control of cryptographic software (summarized in Table 4-3) date back to the days of COCOM (Coordinating Committee for Multilateral Export Controls), an international organization created to control the export and spread of military and dual-use products and technical data. COCOM’s 17 member countries included Australia, Belgium, Canada, Denmark, France, Germany, Greece, Italy, Japan, Luxemburg, The Netherlands, Norway, Portugal, Spain, Turkey, the United Kingdom, and the United States. Cooperating members included Austria, Finland, Hungary, Ireland, New Zealand, Poland, Singapore, Slovakia, South Korea, Sweden, Switzerland, and Taiwan.
In 1991, COCOM realized the difficulty of controlling the export of cryptographic software at a time when programs implementing strong cryptographic algorithms were increasingly being sold on the shelves of stores in the United States, Europe, and Asia. As a result, the organization decided to allow the export of mass-market cryptography software and public domain software. COCOM’s pronouncements did not have the force of law, but were only recommendations to member countries. Most of COCOM’s member countries followed COCOM’s regulations, but the United States did not. COCOM was dissolved in March 1994.
In 1995, negotiations started for the follow-up control regime—the Wassenaar Arrangement on Export Controls for Conventional Arms and Dual-Use Goods and Technologies. The Wassenaar treaty was signed in July 1996 by 31 countries: Argentina, Australia, Austria, Belgium, Canada, the Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Japan, Luxembourg, the Netherlands, New Zealand, Norway, Poland, Portugal, the Republic of Korea, Romania, the Russian Federation, the Slovak Republic, Spain, Sweden, Switzerland, Turkey, the United Kingdom, and the United States. Bulgaria and Ukraine later joined the Arrangement as well. The initial Wassenaar Arrangement mirrored COCOM’s exemptions for mass-market and public-domain encryption software. In December 1998 the restrictions were broadened.
The Council of Europe is a 41-member intergovernmental organization that deals with policy matters in Europe not directly applicable to national law. On September 11, 1995, the Council of Europe adopted Recommendation R (95) 13 Concerning Problems of Criminal Procedure Law Connected with Information Technology, which stated, in part, that “measures should be considered to minimize the negative effects of the use of cryptography on the investigation of criminal offenses, without affecting its legitimate use more than is strictly necessary.”
The European Union largely deregulated the transport of cryptographic items to other EU countries on September 29, 2000, with Council Regulation (EC) No 1334/2000 setting up a community regime for the control of exports of dual-use items and technology (Official Journal L159, 30.1.2000), although some items, such as cryptanalysis devices, are still regulated. Exports to Australia, Canada, the Czech Republic, Hungary, Japan, New Zealand, Norway, Poland, Switzerland, and the United States, may require a Community General Export Authorization (CGEA). Export to other countries may require a General National License for export to a specific country.
Table 4-3. International agreements on cryptography
COCOM (Coordinating Committees for Multilateral Export Controls)
Eased restrictions on cryptography to allow export of mass-market and public-domain cryptographic software.
Wassenaar Arrangement on Export Controls for Conventional Arms and Dual-Use Goods and Technologies
Allows export of mass-market computer software and public-domain software. Other provisions allow export of all products that use encryption to protect intellectual property (such as copy protection systems).
Council of Europe
Recommends that “measures should be considered to minimize the negative effects of the use of cryptography on investigations of criminal offenses, without affecting its legitimate use more than is strictly necessary.”
Export to other EU countries is largely unrestricted. Export to other countries may require a Community General Export Authorization (CGEA) or a General National License.
Table 4-4 summarizes national restrictions on the import, export, and use of cryptography throughout the world as of March 2001.
Table 4-4. National restrictions on cryptography
Domestic use restrictions
Export regulated in accordance with Wassenaar. Exemptions for public domain software and personal-use. Approval is also required for software that does not contain cryptography but includes a plug-in interface for adding cryptography.
Follows EU regulations and Wassenaar Arrangement.
Laws forbid encrypting international radio transmissions of corporations and organizations.
Import and export requires license.
A license is required for design, production, sale, repair, and operation of cryptography.
Requires license for exporting outside of the Benelux.
None currently, although regulations are under consideration.
None currently, but export and import may be regulated by the Myanmar Computer Science Development Council.
Use of cryptography may require license.
Follows pre-December 1998 Wassenaar regulations. Public domain and mass-market software can be freely exported.
People’s Republic of China
Requires license by State Encryption Management Commission for hardware or software where the encryption is a core function.
Use of products for which cryptography is a core function is restricted to specific products using preapproved algorithms and key lengths.
Import allowed if the product is not to be used “for production, development, collection or use of nuclear, chemical or biological weapons.” Export controls are not enforced.
Some export controls in accordance with Wassenaar.
Importers must be registered.
Export controlled in accordance with Wassenaar.
Export requires license in accordance with EU Recommendation and Wassenaar, although a license is not required for mass-market goods.
Some imports and exports may require a license depending on intended function and key length.
France liberalized its domestic regulations in March 1999, allowing the use of keys up to 128 bits. Work is underway on a new law that will eliminate domestic restrictions on cryptography.
Follows EU regulations and Wassenaar; companies can decide for themselves if a product falls within the mass-market category.
Follows pre-December 1998 Wassenaar.
Hong Kong Special Administrative Region
License required for import and export.
None, although encryption products connected to public telecommunication networks must comply with the relevant Telecommunications Authority’s network connection specifications.
Mirror regulations requiring an import license if an export license is needed from Hungary. Mass-market encryption software is exempted.
License required for import.
No import controls. Export regulated under Wassenaar; no restrictions on the export of software with 64-bit key lengths.
Electronic Commerce Act 2000 gives judges the power to issue search warrants that require decryption.
Import and export require a license from the Director-General of the Ministry of Defense.
Use, manufacture, transport, and distribution of cryptography within Israel requires a license from the Director-General of the Ministry of Defense, although no prosecutions for using unlicensed cryptography are known and many Israeli users apparently do not have licenses.
Follows EU regulations.
Encrypted records must be accessible to the Treasury.
Export regulations mirror pre-December 1998 Wassenaar. Businesses must have approval for export of cryptography orders larger than 50,000 yen.
License from the Committee of National Security required for research, development, manufacture, repair, and sale of cryptographic products.
Mirrors EU regulations.
Follows pre-December 1998 Wassenaar.
None, although search warrants can be issued that require the decryption of encrypted messages.
Import and export requires a license from the Ministry of National Security.
Use requires a license from the Ministry of National Security.
Follows Wassenaar. No license required for export to Belgium or Luxemburg.
Police can order the decryption of encrypted information, but not by the suspect.
Follows Wassenaar. Approval is also required for software that is designed for plug-in cryptography.
Sale and use of encryption requires prior approval.
Follows pre-December 1998 Wassenaar.
No import controls. Exports according to Wassenaar.
License required for import and export.
Licenses required for some uses.
“It is reported that Saudi Arabia prohibits use of encryption, but that this is widely ignored.”
Hardware equipment connected directly to the telecommunications infrastructure requires approval. Police, with the consent of the Public Prosecutor, may compel decryption of encrypted materials.
In accordance with pre-December 1998 Wassenaar.
Import and export controls only apply to military cryptography.
No regulations for commercial use or private organizations. Use by government bodies requires prior approval.
Import of encryption devices forbidden by government policy, not legislation. Import of encryption software is not controlled. Export in accordance with Wassenaar.
Export in accordance with Wassenaar and EU regulations.
Export follows Wassenaar. Export of 128-bit symmetric mass-market software allowed to a list of 61 countries.
Use of equipment to decode encrypted radio and television transmissions is regulated.
Import is not controlled. Export mirrors Wassenaar.
Some uses may be regulated.
Follows pre-December 1998 Wassenaar controls.
No obvious restrictions.
Follows EU and Wassenaar restrictions on export.
Regulation of Investigatory Powers Act 2000 gives the government the power to disclose the content of encrypted data.
United States of America
Few export restrictions.
Import requires license.
 By sheer coincidence, the number 4,405,829 is prime.
 “The Influence of ULTRA in the Second World War,” lecture by Sir Harry Hinsley, Babbage Lecture Theatre, Computer Laboratory, University of Cambridge, Tuesday, 19 October 1993. A transcript of Sir Hinsley’s talk can be found at http://www.cl.cam.ac.uk/Research/Security/Historical/hinsley.html.
 For a more detailed report, see “The Risks of Key Recovery, Key Escrow, and Trusted Third-Party Encryption,” by Hal Abelson, Ross Anderson, Steven M. Bellovin, Josh Benaloh, Matt Blaze, Whitfield Diffie, John Gilmore, Peter G. Neumann, Ronald L. Rivest, Jeffrey I. Schiller, and Bruce Schneier, 27 May 1997 (http://www.cdt.org/crypto/risks98/).
 These countries consisted of Cuba, Iran, Iraq, Libya, North Korea, Sudan, and Syria as of March 2001.
 The actual regulations are somewhat more complex than can be detailed in this one paragraph. For a legal opinion on the applicability of the current export regulations to a particular product or software package, you should consult an attorney knowledgeable in import/export law and the Defense Trade Regulations.