Show HN: A key-value store built with homomorphic encryption

#1

Post by akane » Thu, Dec 10, 2020, 8:06 PM UTC

Show HN: A key-value store built with homomorphic encryption
github.com

Re: Show HN: A key-value store built with homomorphic encryption

#2

Post by gigatexal » Thu, Dec 10, 2020, 8:21 PM UTC

Should not be considered secure, only for research. Dang it. The premise sounded really cool!

Re: Show HN: A key-value store built with homomorphic encryption

#3

Post by a3camero » Fri, Dec 11, 2020, 2:42 AM UTC

Neat idea.

Re: Show HN: A key-value store built with homomorphic encryption

#4

Post by oersted » Fri, Dec 11, 2020, 10:49 AM UTC

O(N) to get a key-value (N being the total number of keys) kind of defeats the purpose of a database.

Couldn't you just use a file with (normally) encrypted keys and values? I'm probably slightly misunderstanding the premise of this project.

Re: Show HN: A key-value store built with homomorphic encryption

#5

Post by ficklepickle » Fri, Dec 11, 2020, 1:28 PM UTC

oersted wrote: post #4

O(N) to get a key-value (N being the total number of keys) kind of defeats the purpose of a database. Couldn't you just use a file with (normally) encrypted keys and values? I'm probably slightly misunderstanding the premise of this project.

Non-cryptographer here, but I think homomorphic encryption is fascinating. IIRC, it allows a user to perform operations on encrypted data, without being able to decrypt the original data. I see it as the holy grail of data anonymization and I understand it to be cutting edge stuff.

I'm sure someone will correct me soon :)

Re: Show HN: A key-value store built with homomorphic encryption

#6

Post by ficklepickle » Fri, Dec 11, 2020, 1:41 PM UTC

gigatexal wrote: post #2

Should not be considered secure, only for research. Dang it. The premise sounded really cool!

Here are details about security from HElib[0], the library used by this project.

The premise of HE is really cool. I think it's just early days yet.

[0] https://github.com/homenc/HElib/blob/master/CKKS-security.md

Re: Show HN: A key-value store built with homomorphic encryption

#7

Post by annguyenfoss » Sat, Dec 12, 2020, 1:02 AM UTC

Ankane should stop building non-sense, half-assed open source tools that serve no good purpose..

Re: Show HN: A key-value store built with homomorphic encryption

#8

Post by alexmingoia » Sun, Dec 13, 2020, 4:03 AM UTC

What unique abilities does homomorphic encryption provide here?

Ciphertext can be used for keys or values in a normal key-value store...

Re: Show HN: A key-value store built with homomorphic encryption

#9

Post by akane » Mon, Dec 14, 2020, 5:05 AM UTC

oersted wrote: post #4

O(N) to get a key-value (N being the total number of keys) kind of defeats the purpose of a database. Couldn't you just use a file with (normally) encrypted keys and values? I'm probably slightly misunderstanding the premise of this project.

I mainly thought it was neat there's a way to do matching on encrypted data without deterministic encryption. It's based on this example - https://github.com/homenc/HElib/tree/master/examples/BGV_cou... - which explains more about how it works and why it's O(N).

Re: Show HN: A key-value store built with homomorphic encryption

#10

Post by akane » Mon, Dec 14, 2020, 5:11 AM UTC

alexmingoia wrote: post #8

What unique abilities does homomorphic encryption provide here? Ciphertext can be used for keys or values in a normal key-value store...

Homomorphic encryption reveals less information to the server about the query than if you were to encrypt keys (with deterministic encryption) and values. It hides the fact that the same query was run multiple times and which key (if any) was matched.