Double-Blind Proof of Existence for Decentralized Identities

Abstract

Decentralized identities return control of identities to the identity owners. Although current work enhances the privacy of these publicly stored identities using encryption and zero-knowledge proofs, decentralized identities can still be abused due to the following problems: • identity holders, e.g., blockchain peers, can profile identity owners by looking at “who is reading which identity data”, and • identity verifiers, e.g., applications and websites, learn private data about owners, like their monetary values and previous transactions during the identity linking. In the worst case scenario, the identity holders and verifiers collaboratively profile users to learn more information. As a practical solution, we introduce the notion of Double Blind Proofs of Existence (DBPoE), which shows that an opened DID is committed in one of the constant-sized multi-generator Pedersen commitments (33 Bytes at 128-bit security), and nothing else. Hence, our DBPoE double-blinds identity holders and identity verifiers to mitigate private information leakage. Equally importantly, our multi-generator commitment-based DBPoE is more resistant to graph analysis than other one-of-many proofs, e.g., ring signatures, which we show mathematically using the maximal flow problem. Our DBPoE protocol has a size complexity of $O(\log _{2}(N) + m)$ when the real commitment is hidden in $N$ commitments and $m$ generators are used, e.g., when $m=4$ , a DBPoE of 1000 commitments is only 3 KB.

Keywords

• Decentralized identities
• blockchain
• privacy-preserving cryptography