DAVC: A Verifiable Credentials Model for Dynamic Attribute Management

Abstract

In decentralized ecosystems, Decentralized Identifiers (DID) and Verifiable Credentials (VC) enable self-sovereign identity, cross-domain interoperability, and privacy-preserving data exchange. However, current VC models face critical limitations, including static attribute binding, inefficient updates, high on-chain verification costs, and privacy leakage. To address these issues, this paper proposes a Dynamic Attribute-oriented Verifiable Credential (DAVC) model, designed to support flexible attribute lifecycle management. The model adopts a three-layer architecture that combines minimal on-chain anchoring, off-chain attribute decoupling, and hierarchical recursive verification to enable efficient and scalable identity verification. First, the onchain layer introduces Sparse Merkle Tree (SMT) root hashes to reduce the need for recording off-chain attribute statuses. Second, the off-chain layer achieves semantic isolation between attributes and identities through an Anonymous Attribute Identifier (AID) mechanism, while improving update efficiency via path caching and incremental strategies. Finally, a Hierarchical Recursive Zero-Knowledge Proof (HR-ZKP) mechanism, based on the Halo2 framework, achieves logarithmic complexity in multiattribute proof generation, supporting attribute-level minimal disclosure and structural anonymization. Experimental results demonstrate that DAVC maintains constant on-chain storage, significantly reduces gas consumption, keeps proof sizes within reasonable limits (e.g., 2.9KB for 10 attributes), and achieves proof generation delays within hundreds of milliseconds. Overall system performance exhibits logarithmic growth as the number of attributes increases. The DAVC model achieves a balance between strong privacy protection, high composability, and dynamic identity expression through minimal on-chain data usage and closed verification paths. This provides a valuable reference for composable identity authentication in Web3 scenarios.