Lifecycle of zk-SNARKs
This section explores the mechanisms for tokenizing datasets within the Data Marketplace, converting raw data into blockchain-native, tradeable assets. The approach integrates Proof of Space (PoSp) for decentralized storage security, the InterPlanetary File System (IPFS) for data distribution, and Zero Knowledge Proofs (ZKPs) for privacy-preserving verification.

Trusted Setup Phase
The lifecycle begins with a trusted setup, where a Common Reference String (CRS) is generated through a multi-party computation (MPC) ceremony involving a minimum of 20 participants as specified in the base layer, achieving a collusion risk below 2^(-128). The security of this process relies on the assumption that at least one participant is honest and properly discards their secret contribution. The ceremony coordination is managed through governance mechanisms, ensuring transparency and community oversight.

Off-Chain Verified Proof Generation
Decentralized Computation for Scalable, Secure, and Incentivized Verification
Proof generation occurs off-chain and is performed by specialized entities called Proof Pods, which are part of the marketplace's dedicated infrastructure coordinated through off-chain workers. For a standard 10,000-gate circuit, proof generation requires approximately 10 seconds on standard hardware as specified in the base layer. The computational complexity of this process scales with circuit size and structure, making off-chain execution necessary for complex operations.
In the marketplace ecosystem, Proof Pods serve as dedicated computational resources that generate zero-knowledge proofs for various operations including access control verification, dataset attribute validation, and governance participation.

Optimize Workflows with Decoupled Computation
Users requesting operations that require ZKP verification (such as dataset access) send their requests to the Proof Pod network, which then generates the necessary proofs using circuits like AccessVerifier. These Proof Pods stake ZKP Coins as collateral through staking mechanisms to ensure reliable service and are incentivized through rewards derived from marketplace fees.
This separation of proof generation (handled by Proof Pods) from verification (performed on-chain through the verification infrastructure) addresses the computational asymmetry inherent in zero-knowledge systems, where proof generation is significantly more resource-intensive than verification. By delegating the computational burden to a specialized off-chain network of Proof Pods, the marketplace ensures efficiency while maintaining the security guarantees of the underlying cryptographic protocols.
Verification Phase with Proof of Intelligence (PoI) Integration
Verification is designed to take place on-chain, with an estimated cost of 200,000 gas equivalent weight (consistent with base layer specifications) and utilizing EVM pallet precompiles and native verification pallets for efficiency. The Data Marketplace leverages the base layer's Proof of Intelligence (PoI) framework in two distinct ways:
ZKP Verification as PoI Task
The act of verifying zero-knowledge proofs is considered a PoI-eligible operation, allowing validators to earn rewards for performing these cryptographic verifications through the hybrid consensus system.
Computation Validation
For more complex operations, additional PoI tasks validate specific mathematical computations related to datasets, such as verifying statistical properties or model outputs.

Verification Through zk-SNARK Integration
Balancing Incentive Alignment and Computational Efficiency in the Modular Architecture
This dual integration ensures that the marketplace's operations align with the PoI incentive structure while maintaining the security guarantees of ZKP verification within the modular architecture.
The verification process is remarkably efficient compared to proof generation, requiring only a few milliseconds of computation. This asymmetry is a key advantage of zk-SNARKs in blockchain environments, allowing for complex computations to be validated on-chain without imposing excessive computational burdens on validators.
