Prof. Adi Shamir warns that the increasing complexity of modern microprocessor chips, like this one, has made the existence of small, undetected errors almost a certainty.We carry out so many of today’s web-based activities – online banking, email, Internet shopping …
Shamir, a professor of applied mathematics at the Weizmann Institute of Science in Rehovot, recently circulated a research note to a group of colleagues warning of the dangers which may be quietly brewing within our hard drives. The increasing complexity of modern microprocessor chips – the latest Intel processor contains upwards of 30 million transistors, for example – has made the existence of small, undetected errors almost a certainty, he pointed out. A single mathematical oversight in the design of a chip, he warns, could make it possible for a hacker to breach the virtual barrier created by encryption algorithms and wreak havoc on a global scale.
Coming from Shamir, the warning is not to be taken lightly. In 2003, he was awarded the prestigious A.M. Turing Award – computer science’s answer to the Nobel Prize – for his groundbreaking work in the field of public key encryption. He was rewarded in particular for the development of what is known as the “RSA algorithm”. Working with Ronald Rivest and Leonard Aldeman at MIT in the late seventies, Shamir – the ‘S’ in the acronym – devised the method as a way to secure digital communications between parties who have not previously been in contact with one another. Software based upon the algorithm, which produces a code so complex that cracking it would take millions of years, is widely used today to protect e-commerce from hackers.
“The remarkable thing about this note is that Adi Shamir is saying that RSA is potentially vulnerable,” Jean-Jacques Quisquater, a cryptographic researcher at the Belgium’s Université Catholique de Louvain, told The New York Times .
Although Shamir’s warning remains ostensibly theoretical, the chances are that the flaw already exists, and is merely waiting to be unearthed by a hacker with sufficient intelligence and tenacity. The research note points out that previous incidents – such as the 1994 discovery of an obscure bug in Intel’s Pentium microprocessor – make the existence of this vulnerability a near-certainty.
“Even if we assume that Intel had learned its lesson and meticulously verified the correctness of its multipliers,” he wrote, “there are many smaller manufacturers of microprocessors who may be less careful with their design.”
The risks involved are considerable. If an intelligence organization or hacker identified a mathematical error in a widely used processor, then security software on a PC could be “trivially broken with a single chosen message.” Such an attack would only require the ability to send a single “poisoned” message to a secured computer, which would then indicate the value of the encryption key used by the security system involved. What this means, said Shamir, is that “millions of PCs can be attacked simultaneously, without having to manipulate the operating environment of each one of them individually.”
Intel, however, has responded with pragmatism, pointing out that such an error rests upon many contingencies. “We appreciate these and we look at everything,” spokesman George Alfs told The New York Times, adding, however, that Shamir’s scenario remains a hypothesis, not a reality.
So what’s the average PC user to do – revert to a pre-tech state of blissful analogue ignorance?
One might take a leaf from the professor’s own book, and stay aware of the dangers of computing – while continuing to enjoy the incredible conveniences offered by the modern high-tech world.
After circulating his research note, Shamir quickly assured colleagues that he had no evidence of an imminent attack – spreading the word, rather tellingly, by email.