Hacker News Reader

Comment by KenoFischer | Hacker News Reader

Last I looked, HElib was one of the slower implementations. That said, even in the best case, HE libraries are so slow, that I don't think "overhead" really captures the paradigm correctly, because the performance is highly dependent on the application. The best case "overhead" (doing some operation that the HE scheme natively supports, e.g. arithmetic over a very specific finite field) is probably 5-6 orders of magnitude.

View on HN

Is there any application that will benefit from FHE? I can't think of anything where taking a hit of at least 5 orders of magnitude is worth it over doing the calculation locally on a trusted CPU.

It's not just about outsourcing compute. There's lots of interesting things you can build on top of an FHE primitive. The linked blog post has the private information retrieval example (the database owner doesn't want to send you the whole database, you don't want the database owner to know what you're looking for). I have a worked example for machine learning here: https://juliacomputing.com/blog/2019/11/22/encrypted-machine... (The model owner doesn't want to share the model, you don't want to share the image you're asking them to process).

What's the advantage of using FHE in this example as opposed to Secure MPC?

The communications latency of MPC would probably exceed the time required for FHE in this particular example.

Was just about to post this here if no one else had yet. Really cool stuff! This was the first example I saw of FHE where I was actually able to easily run the underlying demo code [1] on my own system.

[1] https://github.com/JuliaCrypto/ToyFHE.jl

NTT(Japanese ex-national telco, privatized) used to be active on it, because national telco not being able to break citizens’ privacy is good for democracy, should the government or the corporate become corrupt.

Reply on news.ycombinator.com