Published:
An American company is developing a secure operating system that “is optimized for embedded devices running machine learning applications.”
Google announced that it is developing a new operating system called KataOS which, according to its designers, will become a secure platform for machine learning tools and will help protect personal data across different devices.
In a statement, the company states that the Google Research team decided to resolve the problem of digital security by creating “a provably secure platform that is optimized for embedded devices running machine learning applications”.
Although KataOS is still in the stage of elaboration, the company invited users to collaborate on the platform to help “ build intelligent environmental systems that have built-in security by default”.
The new operating system is available on GitHub, in an open source operating system format for all who are interested. KataOS is based on Microkernel seL4, the security of which “is mathematically proven6349413with guaranteed confidentiality, integrity, and availability.”
At the same time, KataOS is almost entirely implemented in the Rust programming language, which, according to Google, makes it possible to remove entire classes such as single errors and buffer overflows, providing “a solid starting point” for software security To make the operating system even more secure, the company’s engineers are creating a reference implementation called Sparrow, which combines KataOS with a ‘software’ platform. ‘ safe.
