Google's OpenSK Offers An Open-Source Rust-Written Security Key Implementation

[phoronix]

Phoronix: Google's OpenSK Offers An Open-Source Rust-Written Security Key Implementation

Google today announced OpenSK as an open-source Rust-based security key implementation supporting FIDO U2F and FIDO2 standards...

http://www.phoronix.com/scan.php?page=news_item&px=Google-OpenSK-Launched

[schmidtbag]

I'm surprised it wasn't written in Go.

[nazar-pc]

I'm surprised it wasn't written in Go.

I would be extremely surprised if it was written in Go.

[bachchain]

The Rust code is written the Tock embedded operating system.

I think there's supposed to be a preposition in there somewhere

[bug77]

I'm surprised it wasn't written in Go.

They probably needed low-level access to some AES platform specific instructions. Plus, I did some hobby benchmarking once, Rust came up as 10x faster in number crunching.

[zamadatix]

They probably needed low-level access to some AES platform specific instructions. Plus, I did some hobby benchmarking once, Rust came up as 10x faster in number crunching.

You can do that in Go as well (or use the AES package in the stdlib). 10x faster in number crunching points to a bad comparison not a bad language, even JavaScript shouldn't be 10x slower in number crunching and it runs in a VM without proper native types.

The choice is almost certainly the direct result of Tock OS being written in Rust and providing a first class Rust userland library (libtock-rs) which replaces the need to port the stdlib of the chosen language for the app (note OpenSK is written with nostd).

[xinorom]

Let me know when they re-implement it in a proper language...

[reavertm]

Let me know when they re-implement it in a proper language...

Not sure what you mean by proper language, doesn't Nitrokey already meet that requirement?

[bachchain]

Let me know when they re-implement it in a proper language...

Java or bust

[xinorom]

Not sure what you mean by proper language

I mean a language that won't be just another dead research project 10 years from now.