Hard for me to understand, but my takeaway is that this technology still requires lots of memory and lots of CPU power. Changing a single encrypted bit still takes on the order of “seconds”.
It’s worth pursuing, because the final result is cool: Ability to change encrypted data in the cloud, without the cloud server ever having access to the encrypted data.
NTRU represents something called post-quantum cryptography (or PQC); these are public key algorithms that are expected (and designed) to work even in the case that quantum computing “destroys” current public key ciphers. This is a pretty interesting demonstration of doing something useful using a PQC cipher. NTRU is one of the more promising ciphers in this class.
Homomorphic encryption is an extremely useful tool, should we get get it performant. It would greatly strengthen “cloud computing”, or any case where a user doesn’t store their own data. Particularly, I think it would interesting to see an email system based on this – the mail provider can still provide service, and users have less concern over their emails being read by the provider.
The problem here is, of course, performance. The paper is notable in that researchers are making promising strides in both of the above areas.
Hard for me to understand, but my takeaway is that this technology still requires lots of memory and lots of CPU power. Changing a single encrypted bit still takes on the order of “seconds”.
It’s worth pursuing, because the final result is cool: Ability to change encrypted data in the cloud, without the cloud server ever having access to the encrypted data.
There are two Big Deal parts to this:
NTRU represents something called post-quantum cryptography (or PQC); these are public key algorithms that are expected (and designed) to work even in the case that quantum computing “destroys” current public key ciphers. This is a pretty interesting demonstration of doing something useful using a PQC cipher. NTRU is one of the more promising ciphers in this class.
Homomorphic encryption is an extremely useful tool, should we get get it performant. It would greatly strengthen “cloud computing”, or any case where a user doesn’t store their own data. Particularly, I think it would interesting to see an email system based on this – the mail provider can still provide service, and users have less concern over their emails being read by the provider.
The problem here is, of course, performance. The paper is notable in that researchers are making promising strides in both of the above areas.