Can Japan Become a World Leader in Encryption? [Continued]

Return to previous page

SETting a new standard

In the current global encryption market, the leader in development of elliptic curve cryptosystems is a Canadian company, Certicom, followed by several Japanese companies. Certicom was founded in 1985 by professors of the University of Waterloo as an incubator entity for the university's encryption development team, the Data Encryption Group.

Certicom today provides its elliptic curve cryptosystem products to such major companies as Motorola, Sterling Commerce, VeriFone, and Atalla. And to promote the technology, Certicom has been involved in standardization activities with organizations such as the Institute of Electrical and Electronics Engineers (IEEE), American National Standards Institute (ANSI), International Standards Organization (ISO), International Electrotechnical Commission (IEC), National Institute of Standards and Technology (NIST), and Internet Engineering Task Force.

Certicom has proposed its elliptic curve cryptosystem to MasterCard International for inclusion in version 2.0 of the SET (Secure Electronic Transaction) protocol. (MasterCard reportedly backs changing to an elliptic curve cryptosystem while Visa continues to support RSA.) Hitachi has similarly been promoting the use of its Keymate/Crypto to MasterCard since late last year. "We are scheduled to demonstrate our product to MasterCard this spring," reveals Chisato Konno, engineering manager of the system product planning department at Hitachi.

Hitachi has run into a roadblock for promoting Keymate/Crypto in foreign markets, however, in the form of some unexpected "administrative guidance" from the Ministry of International Trade and Industry (MITI). MITI has "instructed" Hitachi not to aggressively promote its made-in-Japan encryption technology overseas for the time being.

Governments get into the act

The genesis of this request was complaints faced by the Japanese government from other countries, particularly the US, in a 1996 meeting of the OECD (Organization for Economic Cooperation and Development). The issue in question was the increasing number of encryption products that Japan had been exporting. In the same year, too, a made-in-Japan encryption product had become the topic of fierce discussion between American industry leaders and the US government in light of existing US export controls on encryption products.

An NTT subsidiary, NTT Electronics Technology (NEL), had tried to ship to the US an RSA-based chip with a 1,024-bit key length. Since NEL had developed the chip with an RSA algorithm that had been opened to the public, the development itself was not a problem. The issue was that the key length of the NEL chip was much stronger than that of exportable US chips. The US government reportedly put pressure on the Japanese government over the issue of encryption technology exports, and since then MITI has become very cautious about authorizing the export of encryption systems developed in Japan. MITI has given only a few export approvals, other than those for applications to financial institutions, in the past year.

When Hitachi announced development of its Keymate/Crypto, therefore, MITI advised the manufacturer not to target foreign sales, and even asked Hitachi not to heavily publicize its development of an elliptic curve cryptosystem. Hitachi has followed MITI's guidance, "considering our responsibilities as a company in Japan." As a result, Hitachi's release of the Keymate/Crypto product was combined with that of other intranet security server software. Further, public announcement of the Keymate/Crypto system was placed inconspicuously at the very end of the press release, even though the most newsworthy topic of the entire release was the development of Japan's first commercial elliptic curve cryptosystem product.

An about-face

The new, restrictive policy is an abrupt about-face for MITI, which was originally a strong promoter of developing made-in-Japan encryption technologies. For Hitachi's development of the elliptic curve cryptosystem that is the basis of its Keymate/Crypto, MITI had provided a subsidy through the Information-technology Promotion Agency (IPA), a MITI-affiliated organization, as part of the IPA's "productive software development" projects.

MITI had even envisioned promoting a Japanese version of SET, called SECE (Secure Electronic Commerce Environment). The SECE protocol was developed jointly by Hitachi, NEC, and Fujitsu under the auspices of MITI. The ministry further expected that a made-in-Japan elliptic curve cryptosystem would be incorporated into version 2.0 of SECE. (The first version of SECE was developed by using RSA.)

However, MITI has reversed its attitude under intense gaiatsu (external pressure) from abroad, and stopped backing SECE development. Without MITI's strong guidance, the three SECE promoters have begun experiencing internal discord regarding just whose elliptic curve cryptosystem would be adopted for SECE version 2.0. Even the meaningfulness of SECE itself is now being called into question.

Hitachi's Konno says that the three companies are still discussing the possibility of using an elliptic curve cryptosystem in the next version SECE. For now, though, it seems quite unlikely that they will reach a consensus. Many in the industry now consider even the concept of Japan-focused SECE itself to be impractical.

Who'll take the lead?

The advancement of network technology requires a global standard to ensure interoperability among the systems to be used in the market. To promote a made-in-Japan encryption technology in such a situation, Japanese companies will require strong government backing to make a successful debut on the world stage - something that is lacking right now.

Are Japanese companies too spoiled, and incapable of competing with strong foreign competitors like RSA and Certicom on a global scale? Does the Japanese government lack the foresight, and the courage, to lead Japan forward in encryption technology? Each side has been blaming the other so far, and no clear direction has been shown.

And without direction, Japan stands to fall behind in yet another promising and essential high-tech industry.

Elliptic curves: Why so difficult to decipher?

One reason that elliptic curve technology is so secure is that the underlying theory contains many uncertainty factors. This makes attacks on the elliptic curve cryptosystem very difficult, in spite of the relatively short key length that is employed.

A short key length is a great advantage for commercial applications; it makes the EDLP encryption and decryption processes much faster than, yet just as secure as, the RSA public key cryptosystem.

RSA, for example, which uses DLP logic rather than EDLP, requires a 1,024-bit key length for the SET (Secure Electronic Transaction) protocol. An elliptic curve cryptosystem, in contrast, requires only a 160-bit key length to achieve the same level of security. The elliptic curve cryptosystem is thus seen as suited to implementation in the compact media (such as IC cards) that will be used in tomorrow's electronic commerce applications.



Return to previous page



Back to the table of contents