Tangram: Encryption-friendly SNARK framework under Pedersen committed engines

초록

SNARKs are frequently used to prove encryption, yet the circuit size often becomes large due to the intricate operations inherent in encryption. It entails considerable computational overhead for a prover and can also lead to an increase in the size of the public parameters (e.g., evaluation key). We propose an encryption-friendly SNARK framework, Tangram, which allows anyone to construct a system by using their desired encryption and proof system. Our approach revises existing encryption schemes to produce Pedersen-like ciphertext, including identity-based, hierarchical identity-based, and attribute-based encryption. Afterward, to prove the knowledge of the encryption, we utilize a modular manner of commit-and-prove SNARKs, which uses commitment as a ‘bridge’. With our framework, one can prove encryption significantly faster than proving the whole encryption within the circuit. We implement various Tangram gadgets and evaluate their performance. Our results show 12x - 3500x times better performance than encryption-in-the-circuit.

키워드