Abstract
Smart grids generate sensitive electricity consumption data that requires fine-grained access control for secure sharing with third-party markets. While Attribute-Based Encryption (ABE) offers a solution, traditional implementations suffer from high computational costs during attribute revocation and policy updates. This paper proposes FAC, a fine-grained access control scheme that leverages a Third-party Auditor (TPA) and cloud outsourcing to achieve efficient revocation and policy updates. By decoupling revocation from re-encryption and offloading heavy computations, FAC ensures constant revocation costs and significantly faster decryption compared to existing schemes.
Introduction
The modernization of power systems into smart grids enables bidirectional flows of electricity and information. As shown in Figure 1, the network connects bulk generation to end-users through transmission and distribution layers, with a central control center managing the system. The resulting data is highly valuable for demand-response analysis and targeted services but poses significant privacy risks.
While Attribute-Based Encryption (ABE) is the gold standard for "fine-grained" access control, it has two major performance bottlenecks:
- Attribute Revocation: Traditional ABE requires re-encrypting all existing data when a user's role changes, which is computationally expensive.
- Policy Updating: Changing an access policy usually requires the data owner to download, decrypt, and re-encrypt data under the new policy, wasting bandwidth and CPU.
Methods and System Model
The FAC scheme is built upon Ciphertext-Policy Attribute-Based Encryption (CP-ABE) and Linear Secret-Sharing Schemes (LSSS). The system architecture, illustrated in Figure 2, introduces several key entities: the Control Center (CC), Cloud Server (CS), Third-party Auditor (TPA), Certificate Authority (CA), and Attribute Authorities (AAs).
Core Innovations
- TPA-Assisted Revocation: FAC uses a "Market Version Number" to manage access. Instead of re-encrypting data, the TPA verifies a user's version before helping them generate a decryption token. Revocation simply involves updating the version number, making old keys obsolete without modifying stored ciphertexts.
- Outsourced Policy Updating: To update an access policy, the CC generates a small "update key" for the Cloud Server. The CS then transforms the ciphertext to match the new policy without ever accessing the underlying plaintext.
- Token-Based Decryption: Heavy bilinear pairing operations are offloaded to the TPA. This allows resource-constrained entities (the "Market") to decrypt data with a single exponentiation.
Performance Evaluation
The scheme was evaluated using cryptographic benchmarking with the MIRACL library on a 2.53GHz processor. The authors measured execution time and communication overhead across varying numbers of attributes (up to 100).
- Constant Revocation Cost: In previous state-of-the-art schemes, the cost of revocation increased linearly with the number of ciphertexts. In FAC, the communication overhead for revocation remains constant regardless of the volume of stored files.
- Faster Decryption: FAC significantly outperforms existing schemes in decryption time. For a policy with 100 attributes, FAC takes roughly 500ms, whereas comparison schemes exceed 2000ms.
- Efficient Updates: The policy updating algorithm is significantly faster than standard re-encryption, particularly when the new policy shares attributes with the original one.
Conclusion
FAC addresses the scalability issues of ABE in smart grids by moving heavy computations to the cloud and TPA. By decoupling revocation from re-encryption, it turns a linear-time problem into a constant-time one for storage providers. While the scheme assumes a "curious but honest" model for the Cloud Server and TPA, it provides a robust framework for secure, real-time data sharing. Future work will explore privacy-preserving data aggregation to protect data even before it reaches the Control Center.
Summary updated