zkSNARK library implementation in Go
Succinct Non-Interactive Zero Knowledge for a von Neumann Architecture
, Eli Ben-Sasson, Alessandro Chiesa, Eran Tromer, Madars Virza https://eprint.iacr.org/2013/879.pdfPinocchio: Nearly practical verifiable computation
, Bryan Parno, Craig Gentry, Jon Howell, Mariana Raykova https://eprint.iacr.org/2013/279.pdfOn the Size of Pairing-based Non-interactive Arguments
, Jens Groth https://eprint.iacr.org/2016/260.pdf
Implementation of the zkSNARK Pinocchio protocol and Groth16 protocol from scratch in Go to understand the concepts. Do not use in production.
Not finished, implementing this in my free time to understand it better, so I don't have much time.
Currently allows to do the complete path with Pinocchio protocol and Groth16 protocol :
- write circuit
- compile circuit
- generate trusted setup
- calculate witness
- generate proofs
- verify proofs
Minimal complete flow implementation:
- Finite Fields (1, 2, 6, 12) operations
- G1 and G2 curve operations
- BN128 Pairing
- circuit flat code compiler
- circuit to R1CS
- polynomial operations
- R1CS to QAP
- generate trusted setup
- generate proofs
- verify proofs with BN128 pairing
Improvements from the minimal implementation:
- allow to call functions in circuits language
- allow
import
in circuits language - allow
for
in circuits language - move witness values calculation outside the setup phase
- Groth16
- multiple optimizations
- wasm proof generation
- wasm proof verification
Ongoing experimentation with go-snark compiled to wasm: https://github.com/arnaucube/go-snark/tree/master/wasm
- zkSnark
- zkSnark Groth16
- bn128 (more details: https://github.com/arnaucube/go-snark/tree/master/bn128)
- Finite Fields operations
- R1CS to QAP (more details: https://github.com/arnaucube/go-snark/tree/master/r1csqap)
- Circuit Compiler
The cli still needs some improvements, such as seting input files, etc.
In this example we will follow the equation example from Vitalik's article: y = x^3 + x + 5
, where y==35
and x==3
. So we want to prove that we know a secret x
such as the result of the equation is 35
.
Having a circuit file test.circuit
:
func exp3(private a):
b = a * a
c = a * b
return c
func main(private s0, public s1):
s3 = exp3(s0)
s4 = s3 + s0
s5 = s4 + 5
equals(s1, s5)
out = 1 * 1
And a private inputs file privateInputs.json
[
3
]
And a public inputs file publicInputs.json
[
35
]
In the command line, execute:
> ./go-snark-cli compile test.circuit
If you want to have the wasm input ready also, add the flag wasm
> ./go-snark-cli compile test.circuit wasm
This will output the compiledcircuit.json
file.
Having the compiledcircuit.json
, now we can generate the TrustedSetup
:
> ./go-snark-cli trustedsetup
This will create the file trustedsetup.json
with the TrustedSetup data, and also a toxic.json
file, with the parameters to delete from the Trusted Setup
.
If you want to have the wasm input ready also, add the flag wasm
> ./go-snark-cli trustedsetup wasm
Assumming that we have the compiledcircuit.json
, trustedsetup.json
, privateInputs.json
and the publicInputs.json
we can now generate the Proofs
with the following command:
> ./go-snark-cli genproofs
This will store the file proofs.json
, that contains all the SNARK proofs.
Having the proofs.json
, compiledcircuit.json
, trustedsetup.json
publicInputs.json
files, we can now verify the Pairings
of the proofs, in order to verify the proofs.
> ./go-snark-cli verify
This will return a true
if the proofs are verified, or a false
if the proofs are not verified.
All this process can be done using Groth16 protocol protocol:
> ./go-snark-cli compile test.circuit
> ./go-snark-cli groth16 trustedsetup
> ./go-snark-cli groth16 genproofs
> ./go-snark-cli verify
Example:
// compile circuit and get the R1CS
flatCode := `
func exp3(private a):
b = a * a
c = a * b
return c
func main(private s0, public s1):
s3 = exp3(s0)
s4 = s3 + s0
s5 = s4 + 5
equals(s1, s5)
out = 1 * 1
`
// parse the code
parser := circuitcompiler.NewParser(strings.NewReader(flatCode))
circuit, err := parser.Parse()
assert.Nil(t, err)
fmt.Println(circuit)
b3 := big.NewInt(int64(3))
privateInputs := []*big.Int{b3}
b35 := big.NewInt(int64(35))
publicSignals := []*big.Int{b35}
// witness
w, err := circuit.CalculateWitness(privateInputs, publicSignals)
assert.Nil(t, err)
fmt.Println("witness", w)
// now we have the witness:
// w = [1 35 3 9 27 30 35 1]
// flat code to R1CS
fmt.Println("generating R1CS from flat code")
a, b, c := circuit.GenerateR1CS()
/*
now we have the R1CS from the circuit:
a: [[0 0 1 0 0 0 0 0] [0 0 1 0 0 0 0 0] [0 0 1 0 1 0 0 0] [5 0 0 0 0 1 0 0] [0 0 0 0 0 0 1 0] [0 1 0 0 0 0 0 0] [1 0 0 0 0 0 0 0]]
b: [[0 0 1 0 0 0 0 0] [0 0 0 1 0 0 0 0] [1 0 0 0 0 0 0 0] [1 0 0 0 0 0 0 0] [1 0 0 0 0 0 0 0] [1 0 0 0 0 0 0 0] [1 0 0 0 0 0 0 0]]
c: [[0 0 0 1 0 0 0 0] [0 0 0 0 1 0 0 0] [0 0 0 0 0 1 0 0] [0 0 0 0 0 0 1 0] [0 1 0 0 0 0 0 0] [0 0 0 0 0 0 1 0] [0 0 0 0 0 0 0 1]]
*/
alphas, betas, gammas, _ := snark.Utils.PF.R1CSToQAP(a, b, c)
ax, bx, cx, px := Utils.PF.CombinePolynomials(w, alphas, betas, gammas)
// calculate trusted setup
setup, err := GenerateTrustedSetup(len(w), *circuit, alphas, betas, gammas)
hx := Utils.PF.DivisorPolynomial(px, setup.Pk.Z)
proof, err := GenerateProofs(*circuit, setup, w, px)
b35Verif := big.NewInt(int64(35))
publicSignalsVerif := []*big.Int{b35Verif}
assert.True(t, VerifyProof(*circuit, setup, proof, publicSignalsVerif, true))
Verify Proof generated from snarkjs
Is possible with go-snark
to verify proofs generated by snarkjs
Example:
verified, err := VerifyFromCircom("circom-test/verification_key.json", "circom-test/proof.json", "circom-test/public.json")
assert.Nil(t, err)
assert.True(t, verified)
History of versions & tags of this project:
- v0.0.1: zkSnark complete flow working with Pinocchio protocol
- v0.0.2: circuit language improved (allow function calls and file imports)
- v0.0.3: Groth16 zkSnark protocol added
go test ./... -v
For more details and installation instructions see https://github.com/arnaucube/go-snark/tree/master/vim-syntax
Thanks to @jbaylina, @bellesmarta, @adriamb for their explanations that helped to understand this a little bit. Also thanks to @vbuterin for all the published articles explaining the zkSNARKs.