RISC-V anti-quantum encryption chip is expected to provide future-oriented security
Visit: Alibaba Cloud “88 Bang Bang Festival”: free access to business plan, communication copywriting, and smart accounting tools
It is reported that most of modern cryptography relies on the extreme difficulties that classical computers face when dealing with a large number of mathematical problems such as numbers. But in theory, quantum computers can quickly find answers to problems that classical computers may take hundreds of millions of years to solve.
In order to maintain the leading edge of encryption algorithms relative to the performance of quantum computers, researchers around the world are designing “post-quantum encryption algorithms” that are difficult for traditional and quantum computers to crack.
Georg Sigl, an electrical engineer at the Technical University of Munich, explained that this type of algorithm mostly relies on a lattice-based cryptography and revolves around problems based on multiple points or vectors. In short, a lattice-based encryption algorithm usually selects the target point on which the secret message depends in the lattice, and then adds random noise to make it close to but not completely on some other lattice point.
Without knowing what kind of noise is added, it is very challenging for classical and quantum computers to find the original target point and the corresponding secret information, especially when crystal When the grid is very large.
On the other hand, this encryption algorithm also consumes a lot of computing power when generating randomness and polynomial multiplication. The good news is that Georg Sigl and his colleagues have developed a microchip with a custom accelerator that can perform these steps very efficiently.
Research picture (from: ACS)< /p>
IEEE Spectrum pointed out that the new chip is based on the open source RISC-V standard and complements each other through hardware components and control software to effectively generate randomness and reduce the complexity of polynomial multiplication.
Partners in this work include German industrial giants such as Siemens, Infineon, and Giesecke+Devrient. Take Kyber encryption as an example. Compared with chips based entirely on software solutions, the new chip can speed up about 10 times and consume only 1/8 of the energy.
As early as 2020, the IACR “Transactions on Cryptographic Hardware and Embedded Systems”, the research team has introduced these findings in detail. In addition, this microchip is flexible enough to support another non-lattice-based SIKE post-quantum algorithm.
Kyber is regarded as one of the most promising post-quantum dot matrix cryptographic algorithms, but SIKE needs to consume more computing power. The speed of the new chip is expected to be 21 times that of an encryption chip based on a pure software solution.