Files
paillier/paillier.go
T
Piotr Dyraga 2a33f6e35c Plaintext space check
According to [DJN 10], the plaintext space is `Z_n`. We had this
information in the `Encrypt` and `EncryptWithR` documentation but
sometimes, when encrypting result of more complex equations, it's easy
to miss the fact the result is out of the allowed range.

Here we have just a safety check making sure that if plaintext is out of
allowed space, we report an error.
2018-07-31 16:18:19 +02:00

173 lines
4.8 KiB
Go
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
package paillier
import (
"fmt"
"io"
"math/big"
)
type PublicKey struct {
N *big.Int
}
func (pk *PublicKey) GetNSquare() *big.Int {
return new(big.Int).Mul(pk.N, pk.N)
}
// EncryptWithR encrypts a plaintext into a cypher one with random `r` specified
// in the argument. The plain text must be smaller that N and bigger than or
// equal zero. `r` is the randomness used to encrypt the plaintext. `r` must be
// a random element from a multiplicative group of integers modulo N.
//
// If you don't need to use the specific `r`, you should use the `Encrypt`
// function instead.
//
// m - plaintext to encrypt
// r - randomness used for encryption
// E(m, r) = [(1 + N) r^N] mod N^2
//
// See [KL 08] construction 11.32, page 414.
func (pk *PublicKey) EncryptWithR(m *big.Int, r *big.Int) (*Cypher, error) {
if m.Cmp(ZERO) == -1 || m.Cmp(pk.N) != -1 { // m < 0 || m >= N ?
return nil, fmt.Errorf(
"%v is out of allowed plaintext space [0, %v)",
m,
pk.N,
)
}
nSquare := pk.GetNSquare()
// g is _always_ equal n+1
// Threshold encryption is safe only for g=n+1 choice.
// See [DJN 10], section 5.1
g := new(big.Int).Add(pk.N, big.NewInt(1))
gm := new(big.Int).Exp(g, m, nSquare)
rn := new(big.Int).Exp(r, pk.N, nSquare)
return &Cypher{new(big.Int).Mod(new(big.Int).Mul(rn, gm), nSquare)}, nil
}
// Encrypt a plaintext into a cypher one. The plain text must be smaller that
// N and bigger than or equal zero. random is usually rand.Reader from the
// package crypto/rand.
//
// m - plaintext to encrypt
// E(m, r) = [(1 + N) r^N] mod N^2
//
// See [KL 08] construction 11.32, page 414.
//
// Returns an error if an error has be returned by io.Reader.
func (pk *PublicKey) Encrypt(m *big.Int, random io.Reader) (*Cypher, error) {
r, err := GetRandomNumberInMultiplicativeGroup(pk.N, random)
if err != nil {
return nil, err
}
return pk.EncryptWithR(m, r)
}
// Add takes an arbitrary number of cyphertexts and returns one that encodes
// their sum.
//
// It's possible because Paillier is a homomorphic encryption scheme, where
// the product of two ciphertexts will decrypt to the sum of their corresponding
// plaintexts:
//
// D( (E(m1) * E(m2) mod n^2) ) = m1 + m2 mod n
func (pk *PublicKey) Add(cypher ...*Cypher) *Cypher {
accumulator := big.NewInt(1)
for _, c := range cypher {
accumulator = new(big.Int).Mod(
new(big.Int).Mul(accumulator, c.C),
pk.GetNSquare(),
)
}
return &Cypher{
C: accumulator,
}
}
// Mul returns a product of `cypher` and `scalar` without decrypting `cypher`.
//
// It's possible because Paillier is a homomorphic encryption scheme, where
// an encrypted plaintext `m` raised to an integer `k` will decrypt to the
// product of the plaintext `m` and `k`:
//
// D( E(m)^k mod N^2 ) = km mod N
func (pk *PublicKey) Mul(cypher *Cypher, scalar *big.Int) *Cypher {
return &Cypher{
C: new(big.Int).Exp(cypher.C, scalar, pk.GetNSquare()),
}
}
type PrivateKey struct {
PublicKey
Lambda *big.Int
}
// Decodes ciphertext into a plaintext message.
//
// c - cyphertext to decrypt
// N, lambda - key attributes
//
// D(c) = [ ((c^lambda) mod N^2) - 1) / N ] lambda^-1 mod N
//
// See [KL 08] construction 11.32, page 414.
func (priv *PrivateKey) Decrypt(cypher *Cypher) (msg *big.Int) {
mu := new(big.Int).ModInverse(priv.Lambda, priv.N)
tmp := new(big.Int).Exp(cypher.C, priv.Lambda, priv.GetNSquare())
msg = new(big.Int).Mod(new(big.Int).Mul(L(tmp, priv.N), mu), priv.N)
return
}
type Cypher struct {
C *big.Int
}
func (this *Cypher) String() string {
return fmt.Sprintf("%x", this.C)
}
func L(u, n *big.Int) *big.Int {
t := new(big.Int).Add(u, big.NewInt(-1))
return new(big.Int).Div(t, n)
}
func minusOne(x *big.Int) *big.Int {
return new(big.Int).Add(x, big.NewInt(-1))
}
func computePhi(p, q *big.Int) *big.Int {
return new(big.Int).Mul(minusOne(p), minusOne(q))
}
// CreatePrivateKey generates a Paillier private key accepting two large prime
// numbers of equal length or other such that gcd(pq, (p-1)(q-1)) = 1.
//
// Algorithm is based on approach described in [KL 08], construction 11.32,
// page 414 which is compatible with one described in [DJN 10], section 3.2
// except that instead of generating Lambda private key component from LCM
// of p and q we use Euler's totient function as suggested in [KL 08].
//
// [KL 08]: Jonathan Katz, Yehuda Lindell, (2008)
// Introduction to Modern Cryptography: Principles and Protocols,
// Chapman & Hall/CRC
//
// [DJN 10]: Ivan Damgard, Mads Jurik, Jesper Buus Nielsen, (2010)
// A Generalization of Paillier’s Public-Key System
// with Applications to Electronic Voting
// Aarhus University, Dept. of Computer Science, BRICS
func CreatePrivateKey(p, q *big.Int) *PrivateKey {
n := new(big.Int).Mul(p, q)
lambda := computePhi(p, q)
return &PrivateKey{
PublicKey: PublicKey{
N: n,
},
Lambda: lambda,
}
}