The Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT has announced a newly-developed system designed to keep online communication secure by running them through complex mathematical algorithms.
Called Fiat Cryptography, the code is currently being used to secure about 90 percent of communications sent and received by Google’s Chrome browser.
Google has since used Fiat Cryptography in its open-source cryptographic library, BoringSSL. The library is being used by a multitude of programs including the Chrome browser and numerous Android apps.
CSAIL researchers presented their paper on Fiat Cryptography at the EEE Symposium on Security and Privacy in May. CSAIL graduate student Andres Erbsen was credited as first author while fellow grad students Jade Philipoom and Jason Gross were listed as co-authors. They were joined by Adam Chlipala and Robert Sloan.
Faster than manual encryption
The code can automatically generate and simultaneously verify optimized cryptographic algorithms for all hardware platforms. Perhaps most notably, this process used to be done by hand as recently as just a year ago, as the technology was deployed at MIT labs last year.
Researchers first looked for a solution in C programming and assembly languages and then transferred the best-performing algorithms they identified for each architecture to their own code library.
They then converted programming languages into code using a compiler and proffed them automatically with a mathematical theorem prover called Coq. Each iteration was then tested until the best-performing one is identified for each particular chip architecture.
The researchers found that their automated process was able to match the performance of the best code manually written by humans and was able to complete it much faster.
“It’s basically like taking a process that ran in human brains and understanding it well enough to write code that mimics that process,” said Chlipala, who along with rest of the research team is working on making their compiler find optimized algorithms even faster.
This content was originally published here.