Adaptive-secure identity-based inner-product functional encryption and its leakage-resilience

Linru Zhang, Xiangning Wang, Yuechen Chen, Siu-Ming Yiu

TL;DR

This paper introduces and constructs the first adaptive-secure Identity-Based Inner-Product Functional Encryption (IBIPFE) scheme, which allows users to compute inner products of encrypted data based on their identities. They then extend this to a leakage-resilient version, meaning it can securely function even if an adversary learns some information about the secret keys.

What Problem Does This Paper Address?

Traditional encryption schemes often have an "all-or-nothing" approach: either you can decrypt a message entirely, or you can't decrypt it at all. Functional Encryption (FE) offers a more granular approach, allowing users with specific secret keys to learn only a desired function of the encrypted data. Inner-Product Functional Encryption (IPFE) is a specific type of FE useful for applications like weighted averages.

This paper tackles two key limitations of existing IPFE schemes:

  1. Access Control: How can we ensure that only authorized users can compute the inner product? This requires an identity-based access control mechanism.
  2. Leakage Resilience: In real-world scenarios, adversaries might gain partial information about secret keys (e.g., through side-channel attacks). How can we design FE schemes that remain secure even under such leakage?

The authors address these by proposing Identity-Based Inner-Product Functional Encryption (IBIPFE) and then extending it to be leakage-resilient in the Bounded-Retrieval Model (BRM).

Methods

The paper introduces two main contributions:

  1. Adaptive-Secure IBIPFE:
    • They formalize the definition of IBIPFE, where ciphertexts and secret keys are associated with identities and vectors. Decryption only succeeds if the identities match, and it reveals the inner product of the vectors.
  2. They construct the first adaptive-secure IBIPFE scheme based on the Decisional Bilinear Diffie-Hellman (DBDH) assumption. The "adaptive-secure" aspect means the adversary can make secret key queries at any point during the security game, making it a stronger security notion than selective security.
  3. Leakage-Resilient IBIPFE in the Bounded-Retrieval Model (BRM):
    • They extend their IBIPFE scheme to be leakage-resilient in the BRM. This model allows an adversary to learn a bounded number of bits ($l$) from each secret key.
  4. The construction relies on a novel primitive called an Identity-based Inner-Product Hash Proof System (IBIP-HPS). This primitive is designed to work with inner products and identities, and its decapsulation algorithm is crucial for building the leakage-resilient scheme.
  5. They then build a leakage-smooth IBIP-HPS from their IBIPFE scheme, which is a key step towards achieving leakage resilience.
  6. Finally, they compile this leakage-smooth IBIP-HPS into a leakage-resilient IBIPFE scheme in the BRM. The efficiency of their scheme (except for secret key size) is independent of the leakage bound $l$, as required by the BRM.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Bridging Identity and Functionality: The core idea is to combine identity-based encryption (IBE) principles with inner-product functional encryption. This allows for fine-grained access control directly within the functional encryption framework.
  • The Power of IBIP-HPS: The introduction of IBIP-HPS is a significant technical contribution. It acts as a bridge, enabling the construction of leakage-resilient schemes by providing a structured way to handle secret keys and ciphertexts that is amenable to leakage analysis.
  • Leakage-Smoothness as a Tool: The concept of "leakage-smoothness" for IBIP-HPS is crucial. It ensures that even with partial information about secret keys, the output of the decapsulation algorithm remains sufficiently random or unpredictable, which is essential for proving leakage resilience.
  • Adaptive Security is Harder: Achieving adaptive security for functional encryption is generally more challenging than selective security. The paper demonstrates how to achieve this for IBIPFE, which is a more robust security guarantee.

Key Results

  • The paper presents the first adaptive-secure IBIPFE scheme based on the DBDH assumption.
  • They construct the first leakage-resilient IBIPFE scheme in the Bounded-Retrieval Model (BRM).
  • The proposed leakage-resilient scheme's efficiency parameters (master public key size, master secret key size, ciphertext size, encryption/decryption time) are independent of the leakage bound $l$, meeting the BRM requirements.
  • The leakage ratio of their scheme is shown to be approximately 1 for large leakage parameters.

Limitations and Future Work

  • Discrete Logarithm Requirement: The presented adaptive-secure IBIPFE scheme requires computing discrete logarithms for decryption. The authors note that methods exist to reduce this cost, referencing prior work.
  • Efficiency of IBIP-HPS: While the construction is proven secure, the efficiency of the IBIP-HPS construction and its subsequent compilation into a leakage-resilient IBIPFE might be a point for future optimization.
  • Specific Leakage Models: The work focuses on the Bounded-Retrieval Model. Future work could explore leakage resilience in other models or with different leakage functions.
  • Practicality: As with many advanced cryptographic primitives, the practical efficiency and overhead of these schemes would need further investigation for real-world deployment.

Summary updated

Abstract

There are lots of applications of inner-product functional encryption (IPFE). In this paper, we consider two important extensions of it. One is to enhance IPFE with access control such that only users with a pre-defined identity are allowed to compute the inner product, referred as identity-based inner-product functional encryption (IBIPFE). We formalize the definition of IBIPFE, and propose the first adaptive-secure IBIPFE scheme from Decisional Bilinear Diffie-Hellman (DBDH) assumption. In an IBIPFE scheme, the ciphertext is related to a vector $\vec{x}$ and a new parameter, identity ID. Each secret key is also related to a vector $\vec{y}$ and an identity ID'. The decryption algorithm will output the inner-product value $<\vec{x}, \vec{y}>$ only if ID $=$ ID'. The other extension is to make IBIPFE leakage resilient. We consider the bounded-retrieval model (BRM) in which an adversary can learn at most $l$ bits information from each secret key. Here, $l$ is the leakage bound determined by some external parameters, and it can be set arbitrarily large. After giving the security definition of leakage-resilient IBIPFE, we extend our IBIPFE scheme into a leakage-resilient IBIPFE scheme in the BRM by hash proof system (HPS).
BibTeX
@article{zhang2020adaptive,
  author = {Linru Zhang and Xiangning Wang and Yuechen Chen and Siu-Ming Yiu},
  title = {{Adaptive-secure identity-based inner-product functional encryption and its leakage-resilience}},
  journal = {IACR Cryptol. ePrint Arch.},
  year = {2020},
  doi = {10.1007/978-3-030-65277-7\_30},
}
Figure