Noted Israeli cryptographer Adi Shamir (the "S" in RSA Security and the middle one in the picture at right) has made an obvious, but nonetheless important, observation about the security problems that would ensue should a math error be found in any widely used computer chip.
Such an error could allow intelligence agencies and industrial spies to crack messages protected by public key cryptography. It could also allow hackers to break the protections in software used for e-commerce.
The New York Times's John Markoff says Shamir made the observation in a research note he sent to cryptographer colleagues recently.
Shamir's observation isn't new. A division error found in Intel's Pentium microprocessor in the mid-90s first raised the specter of serious computational problems caused by buggy chips -- though that particular problem wasn't considered widespread and Intel claimed that it would affect spreadsheet users only once every 27,000 years. But given the greater reliance on cryptography these days in regular business correspondence and e-commerce transactions, the kind of bug Shamir describes -- which is only hypothetical at this point -- would be far more serious.
Shamir's note is all the more remarkable, a cryptographer says in Markoff's story, because it suggests that Shamir's own RSA algorithm could be vulnerable.
Shamir notes that laws governing trade secrets prevent users of such chips from being able to verify that the chips were made correctly.
“Even if we assume that Intel had learned its lesson and meticulously verified the correctness of its multipliers,” Shamir wrote in his note, according to the Times, “there are many smaller manufacturers of microprocessors who may be less careful with their design.”
Photo: Ron Rivest, Adi Shamir and Len Adelman/RSA Security