Re: Semantic secure homomorphic encryption

From: Aldar C-F. Chan (aldar_at_comm.utoronto.ca)
Date: 02/23/05 

Date: Wed, 23 Feb 2005 16:35:56 GMT

"David Wagner" <daw@taverner.cs.berkeley.edu> wrote in message
news:cvgq0u$115k$3@agate.berkeley.edu...
> >Besides Paillier's, is it possible to have encruption scheme which is
>>both semantic secure and homomorphic?
>
> Well, sure. You seem to be aware that Paillier's is one example.
> You can trivially generate other examples. For instance, if E_k(x)
> denotes encryption with Paillier's, then I think E'_k(x) = (0,E_k(x))
> also satisfies your requirements.

What's the purpose of adding a "0" to increase the communication
overhead? Perhaps there are so other gains, but I cannot see.

>
> Oh, and El Gamal also satisfies the requirements. There are others
> as well.

El Gamal and Paillier are different kinds of homomorphism.
El Gamal is multiplicative homomorphic, i.e. you can find
E(x,y) from E(x) and E(y); whereas, Paillier
is additive homomorphic where you can find E(x,y).

Quantcast