Hi all! Maybe this is not the right category, and maybe some people would rather not talk about this in public, but I will ask anyway: does anyone know what patents are out there related to FHE?
For instance, the Concrete-library has the BSD3-Clause-Clear license, which explicitly suggests that there may be patents applying to this code, and that a user should ask for a license individually.
Even in the case of OpenFHE, where no such clauses have been added to the license, there could be a third party holding a patent and making a claim afterwards, if someone e.g. uses the OpenFHE code for a commercial project.
Also: do general lattice patents also apply to FHE? Might using NTRU-based FHE in a sense be “safer” than using FHE based on ring-LWE?
Hi @Fmpala. hanks for posting and asking. I’m one of the main people behind the OpenFHE project, and we’re of course all very happy that it is getting attention and uptake.
There are several things that you’re asking about - 1) licenses, 2) patents and IP, and 3) NTRU-based FHE.
In general OpenFHE is intended to provide implementations of all the major FHE schemes for practical and production-ready use, along with building blocks that can be easily modified and integrated. We specifically chose the BSD 2-clause license because it is so open and commercial-friendly. OpenFHE is currently used by several academic, commercial and government projects to support highly secure collaboration on sensitive data. I can’t speak for other projects and why they did or didn’t choose specific license models. We at OpenFHE wanted to use a very common, very well understood and very popular license, which is why we specifically chose 2-clause BSD over the many possible options.
On purpose we make no claims about any patents in the OpenFHE license, primarily because the devilish challenge of IP in general is that it can be so very hard to provide guarantees that no one has ever tried to file a relevant patent on anything relevant to any existing IP. For one, there are a lot of spurious patents that get filed every year, especially related to cryptography in general. Even when a patent is filed it isn’t publicly published for months or years. This of course isn’t a challenge just for OpenFHE or just for open-source FHE or even just a problem for cryptography or open-source, but it is a challenge for anything that might reasonably be patented! For one it is just so hard to prove a negative, and secondly it is easy to make false claims about incorrect patent rights.
I’ll defer to others on answering your question on NTRU-based FHE. It might be better to ask that question in its own thread. The folks who would normally happily address this question might not read through your post to see that question.