Abstract
Identity-based proxy re-encryption (IB-PRE) allows a proxy to convert a ciphertext encrypted under a delegator's identity into one under a delegatee's identity without accessing the underlying message. However, traditional IB-PRE requires exact identity matches, making it unsuitable for "noisy" biometric identities like fingerprints. This paper introduces Fuzzy Identity-Based Proxy Re-Encryption (FIB-PRE), where identities are viewed as sets of attributes. Decryption is possible if the identity used for encryption and the user's private key identity are sufficiently close based on a "set overlap" metric. We formulate a security model for FIB-PRE and present a construction proven to be Chosen-Ciphertext Attack (CCA) secure under the decisional bilinear Diffie-Hellman (DBDH) assumption in the random oracle model.
Introduction
Proxy Re-Encryption (PRE) is a critical primitive for secure data delegation in cloud environments. When integrated with Identity-Based Encryption (IBE), it allows users to use descriptive strings, such as email addresses, as public keys. A significant challenge arises when using biometric data (e.g., fingerprints or iris scans) as identities: biometric samples are inherently "noisy" and never identical across multiple readings. While "Fuzzy" IBE addresses this by allowing decryption based on attribute overlap, existing schemes failed to combine this flexibility with high-level Chosen-Ciphertext Attack (CCA) security and collusion resistance—the ability to prevent a proxy and delegatee from conspiring to recover the delegator's secret key.
Methodology
The proposed FIB-PRE scheme achieves its goals through several key cryptographic techniques:
- Fuzzy Identity Representation: Identities are treated as sets of descriptive attributes. A ciphertext re-encrypted for identity $\omega'$ can be decrypted by a user with identity $\omega$ if and only if $|\omega \cap \omega'| \ge d$, where $d$ is a predefined threshold.
- Fujisaki-Okamoto (F-O) Conversion: To elevate the scheme from passive security (CPA) to active security (CCA), the authors utilize the F-O conversion. This ensures the system remains secure even if an adversary has access to a decryption oracle, providing both IND-level1-CCA and IND-level2-CCA security.
- Token-Controlled Re-Encryption: To ensure collusion resistance, the re-encryption process incorporates a random token $\theta$. The proxy performs a partial decryption that requires the delegatee to use their fuzzy identity to recover $\theta$ before the final message can be accessed, preventing the proxy from gaining unauthorized information.
- Bilinear Pairings: The mathematical foundation relies on bilinear maps over cyclic groups, utilizing Lagrange coefficients to efficiently manage attribute matching during the fuzzy decryption process.
Key Results
The primary contribution of this work is the first FIB-PRE scheme that is both CCA-secure and collusion-resistant. The authors provide a rigorous mathematical proof of security under the Decisional Bilinear Diffie-Hellman (DBDH) assumption in the random oracle model.
Beyond theoretical proofs, the scheme's practical advantages are demonstrated through a comparative analysis with existing Identity-Based Proxy Re-Encryption (IB-PRE) methods.
As shown in the comparison above, the proposed scheme is the only one to simultaneously provide identity fuzziness, CCA security, and collusion resistance. This makes it uniquely suited for real-world applications involving biometric authentication and secure cloud delegation.
Conclusion and Future Work
This research establishes a new benchmark for flexible and secure identity-based delegation. By formalizing the security models for fuzzy proxy re-encryption, the authors provide a framework for future developments in biometric-based cryptography.
Current limitations include the reliance on the random oracle model and the "selective-identity" security proof. Future research directions include:
- Developing a CCA-secure FIB-PRE scheme in the standard model.
- Achieving adaptive-identity security, which offers more robust protection against dynamic adversaries.
- Extending the construction to support multi-hop delegation, allowing data to be securely re-encrypted across multiple users in a chain while maintaining biometric fuzziness.
Summary updated