The University of Texas at Austin recently announced that two of its researchers have developed a new and better method to perform the computations necessary for random number generation. This new method will have important implications for cryptography and, by extension, cybersecurity.
Researchers Make a Splash
A draft paper has been published and will be presented in June at the Symposium on Theory of Computing. The paper is one of three that have been recognized as the best submitted for consideration.
Since it was published for peer review and comment on the Electronic Colloquium on Computational Complexity in August 2015, it has been causing excitement in the math community.
In its official statement, the university quoted Yael Kalai, a senior researcher working in cryptography at Microsoft Research New England. “When I heard about it, I couldn’t sleep,” he said. “I was so excited. I couldn’t believe it. I ran to the (online) archive to look at the paper. It’s really a masterpiece.”
Truly Random Number Generation Could Change Security
Random numbers, used to generate the keys necessary for encryption, are the very basis of information security. If random numbers are not truly random, and can therefore be reproduced, the keys can be reproduced as well.
The paper described a method of combining two “weakly random” number sequences and combining them into one truly random number. It sidesteps previous restrictions on the streams used in computation, requiring fewer computational resources and ultimately resulting in a higher quality of randomness.
The Immediate Impact
The work is theoretical in scope but outlines a way that practical implementations may be pursued. However, SecurityWeek reported that Vincent Rijmen, one of the two developers of the Advanced Encryption Standard (AES), is interested but not terrifically enthusiastic about the project.
It “is probably important within its own context,” he told SecurityWeek, “that is, deep theoretic reflections on randomness and cryptography.”
The opinion that it does not currently hold much practical value within cryptography was also reinforced by professor Ross Anderson of the Cambridge University Computer Laboratory. He told SecurityWeek that it was “unlikely to be of much engineering interest” for the time being.
Despite the need for further development, the concept of using less computationally expensive randomness streams for truly random number generation can only benefit encryption efforts.
Principal, PBC Enterprises