IBM's New Algo

[John Young <jya@pipeline.com>]

   The New York Times, May 7, 1997, p. D5.

   I.B.M. Researchers Develop A New Encryption Formula

   By Laurence Zuckerman

   I.B.M. plans to announce today that two of its researchers
   have come up with a new computer encryption formula that
   they say is nearly impossible to crack.

   The International Business Machines Corporation said that
   the breakthrough was still a long way from being employed
   outside the lab and that it did nothing to resolve the
   running dispute between the computer industry and the
   Federal Government over whether law enforcement agencies
   should be given access to encrypted communications. But it
   could ultimately help reduce the vulnerability of so-called
   public-key encryption, which is the favored security method
   used to safeguard commerce and privacy on the Internet.

   "They've made a big advance," said Joan Feigenbaum, a
   researcher at AT&T Labs in Murray Hill, N.J., who is
   familiar with the work of the two computer scientists who
   developed the system, Miklos Ajtai and Cynthia Dwork of
   I.B.M.'s Almaden Research Center in San Jose, Calif.
   "Scientifically, this is a big step in the right
   direction."

   But Bruce Schneier, a computer security consultant in
   Minneapolis and author of a standard textbook on
   cryptography, dismissed the news.

   "Theoretically it is important, but as a security
   breakthrough there is nothing new," he said.

   In public-key encryption, the sender of an electronic
   communication uses software that automatically scrambles
   the information by incorporating a publicly known numerical
   key. Decoding the scrambled transmission requires a private
   key, a number supposedly known only by the recipient.

   The security of the system depends, among other things, on
   how difficult it is for an electronic eavesdropper to crack
   the code using a powerful computer. If some of the codes
   that are generated by the system are difficult to break but
   others are easy, the system is inherently weak. I.B.M. said
   that its new system was the first to generate hundreds of
   codes at random, each of which is as difficult to crack as
   the hardest instance of the underlying mathematical
   problem.

   The system is based on a problem that has defied solution
   by mathematicians for 150 years, I.B.M. said.

   Mr. Schneier said that the cryptographic formulas now in
   use were already robust enough. The biggest challenge, he
   said, is creating security systems in the real world that
   are not vulnerable to hackers.

   "Cryptography is a lot more than math" he said.

   [End]