The world is preparing for the potential arrival of the quantum computer… and the set of threats that would accompany its advent. But the path to “post-quantum” digital security is fraught with pitfalls, as a last-minute twist in a major cryptography contest has just confirmed.
As Phong Nguyen, cryptanalyst at the National Institute for Research in Digital Sciences and Technologies (Inria) reminds us, “it has been known since the 1990s that a quantum computer might break the main so-called cryptographic systems “public key” », that secure the Internet today. As a reminder, a quantum computer is a machine of which we know the theoretical foundations but which no one has yet managed to build. A kind of overpowered and extremely fast calculator.
“We estimate that, within ten or fifteen years, such a machine might exist”, says Dustin Moody, mathematician at the National Institute of Standards and Technology (NIST), the American organization responsible for defining standards in terms of computer security. An opinion shared, in France, by the National Agency for the Security of Information Systems (Anssi), for which the quantum computer might potentially arrive “during the decade 2030-2040”. « [En 2016]to anticipate this threat, we therefore asked experts from all over the world to design new cryptosystems, which would be resistant to quantum computer attacks.details Dustin Moody who is also responsible for the post-quantum cryptography standardization project for the NIST.
In the shoes of hackers
To assess the quality of the proposed algorithms, this organization asked the community of cryptologists to put themselves in the shoes of hackers by looking for flaws in these schemes. Objective: successively eliminate from the race the algorithms deemed too vulnerable. Of the 69 algorithms selected in 2017 to participate in the competition, seven “finalists” have passed the successive selection stages and are still in the running. NIST has also compiled a kind of supplemental list, “to have fallback solutions if the finalists ultimately turn out to be vulnerable, and to maintain diversity in the proposals we are studying”, says Dustin Moody.
Well took them. Because one of the seven finalists, the scheme called “Rainbow”, has just suffered a serious setback, while the third (and penultimate) round of the competition is due to end at the end of March. Ward Beullens, postdoctoral researcher at the Zurich research laboratory of the IBM company, proposed in a “preprint” dated February 21 an attack on Rainbow that seriously damages the crypto community’s confidence in this algorithm. “If Rainbow were deployed as a cryptographic standard, thanks to the attack [de Ward Beullens] anyone might digitally impersonate someone else”, summarizes Ludovic Perret, cryptologist at Sorbonne University and co-founder of the start-up CryptoNext Security. He is also co-author of an algorithm retained in the complementary list of the NIST.
You have 40.29% of this article left to read. The following is for subscribers only.
