Realising AES-advanced encryption standard. (Security).

Joan Daemon-Proton World

The principle behind the Enigma machine is still used for data encryption today. Every time you enter your PIN when withdrawing cash from an ATM, the data stored on the card is encrypted and sent to the issuer, who decrypts it and then encrypts the reply which is decrypted by the ATM. However, things have advanced since the 1940's: instead of rotating wheels, complicated mathematical calculations, called cryptographic algorithms, are used. In the case of e-purse smart cards, the card and the terminal challenge each other to solve an algorithm, which they both do by providing the "key". Once this mutual authentication has been done, the transaction can proceed.

Most of the cryptographic algorithms used today are based on the Data Encryption Standard (DES) a standard developed by IBM and adopted by the US government's National Institute for Standards & Technology (NIST) for protecting sensitive information in 1977. Once in use by the government, it spread rapidly into the private sector, becoming a de- facto global standard.

Modern computers are, however, infinitely more powerful and fast than they were twenty-three years ago. Some can now "try" all the possible keys for a DES algorithm in two hours, rendering DES insufficiently secure for many transactions. The Security Products Development community is well aware of this, and most secure applications migrated in the 1990s to Triple DES, a system that uses three linked DES algorithms to secure transactions. A typical DES algorithm is 56 bits long, with a possible 720,000,000,000,000,000 keys. Modern computers are developing fast, and it is only a matter of time before machines will exist that can break Triple DES security. Therefore, in 1997 NIST decided to start a quest for a new global data encryption standard, called the Advanced Encryption Standard (AES). It decided to hold a world-wide cryptography "Olympics", inviting the world's cryptographer community to submit new algorithms for evaluation and testing by their peers. Fifteen submissions were made that met the minimum re quirements, which were rigorously tested and "attacked" by teams of cryptographers all over the world. By October 1999, five "finalists" had emerged, all of which had resisted attack and had the potential for widespread use. They were MARS from IBM, RC6 from RSA Security; Twofish from Counterpane; Serpent from a British- Danish-Israeli trio of cryptographers and Rijndael (pronounced Rain-dahl) from two Belgians, Vincent Rijmen and Joan Daemen.

The final stage tested all the algorithms in a variety of environments, to select one that was easy to use, flexible and fast. A year later, in October 2000, NIST announced that Rijndael had been chosen for ABS. This was quite a coup for Rijmen (a 30-year-old post-doctoral researcher at the Katholieke Universiteit Leuven (KUL, Leuven University) and Daemen (a 35-year-old security products development cryptographer at Proton World in Brussels), who developed Rijndael in their spare time, and didn't think they stood a chance against the 11 person, full-time IBM team. "it was like entering a competition to design a PC and discovering that Bill Gates was the competition" said Joan Daemen.

Rijndael supports key sizes of 128, 192 and 256 bits. To illustrate the advance from DES, a 128-bit key size has a possible 340,000,000,000,000,000,000,000,000,000,000,000,000, keys. NIST spokesman Jim Foti said "if a machine existed that could recover a DES key in a second (i.e. try 255 keys per second) it would take 149,000,000 million years to crack a 128-bit AES key. To put this in perspective, the universe is believed to be less than 20,000 million years old." NIST will propose AES for adoption as a Federal Information Processing Standard (FIPS). It is expected to be adopted in mid-2001 for use with unclassified data. Initial estimates are that AES will be used as a FIPS for at least twenty years.

The news about AES comes at a crucial period in the development of e- commerce. As cash is not an option when buying goods and services from websites, credit cards have become a de- facto "standard". But their use relies on cardholders typing in their card numbers, expiry dates, names and addresses and then sending them "in clear" (uncrypted) over the Internet. The message may be intercepted by a hacker or the staff at the webshop may copy it down and use it fraudulently somewhere else. The credit card companies report that e-commerce represents 1 % of their transaction volume, but generates 50% of their disputed payments and charge-backs. Other studies show that around 70% of Internet shoppers do not complete their transactions, in many cases due to a lack of confidence in the security of their card data. Moving to AES-protected smart cards for credit and debit transactions should therefore build consumer confidence and give e- commerce a much-needed boost Joan and Vincent have renounced all Intellectual Pro perty Rights to Rijndael, as required by the NIST rules and plan to continue in their current "day jobs": Vincent at the COSIC Research Laboratory at the KUL and Joan as a senior member of the team of over twenty smart card security experts from several countries, who are working to develop the next generation of Proton technology smart card security modules.

Footnote:

How AES is different from DES.

DES uses 56-bit keys, where as AES uses either 128-bit, 192-bit or 256-bit keys. There are around 72,000,000,000,000,000 possible "combinations" for a DES key, and around 3,400,000,000,000,000,000,000,000,000,000,000,000,000 possible "combinations" for a 128-bit AES key.

The COSIC lab's home page is http://www.esat.kuleuven.ac.be/cosic and a Rijndael page is available at http://www.esat.kuleuven.ac.be/~rijmen/rijndael. There is also more information on NIST's website at http://www.nist.gov/public-affairs/releases/g00-176.htm and some of the cryptographic research work at Proton World can be found at http://www.protonworld.com/research. The article is part of a series for ISSE 2001 Info Security Solutions Europe 26/28 Sept 2001 www.eema.org/isse.