zk-SNARK란 무엇인가

zk-SNARKs fundamentally enable verifiable computation in a zero-knowledge setting. The underlying mathematical framework typically involves mapping the computational problem to a system of polynomial equations, often represented as an arithmetic circuit. The prover constructs a proof demonstrating knowledge of a witness satisfying these equations. Common constructions utilize pairings on elliptic curves to check polynomial evaluations at specific points, forming the basis of the verification process. The succinctness property is achieved by compressing the verification information into a small proof, often through techniques like polynomial commitments and the use of cryptographic accumulators. The non-interactive nature is typically realized using the Fiat-Shamir heuristic, transforming an interactive proof system into a non-interactive one by using a [hash function](/ko/terms/hash-function) to derive the verifier's challenges. A significant trade-off is the requirement for a trusted setup phase in many SNARK constructions (e.g., Groth16), where toxic waste must be securely destroyed. Failure to do so allows the creation of undetectable fake proofs. Newer SNARK variants and alternative zero-knowledge proof systems like zk-STARKs aim to mitigate or eliminate this trusted setup requirement.