TL;DR
This paper introduces VO-MAACS, a novel scheme for secure and verifiable access control in fog-cloud computing environments. It offloads most encryption and decryption tasks to fog devices, allowing for efficient data access while ensuring the integrity of outsourced computations through a verification mechanism. The scheme also incorporates efficient user and attribute revocation, addressing key challenges in dynamic fog-cloud systems.
What Problem Does This Paper Address?
The rapid growth of the Internet of Things (IoT) and big data has led to an explosion in the number of connected devices and the volume of data generated. Fog computing, which extends cloud capabilities to the network edge, offers a promising solution to address the bottlenecks in data transmission and storage. However, this distributed architecture introduces new security and privacy concerns, particularly regarding data access control. Traditional access control methods struggle with the dynamic nature and resource constraints of fog devices. Ciphertext-Policy Attribute-Based Encryption (CP-ABE) is a suitable cryptographic tool for fine-grained access control, but its computational demands are high, especially in multi-authority settings common in fog-cloud environments. Furthermore, outsourcing these computations to less trusted fog devices raises issues of "laziness" (incomplete or incorrect computations) and the need for efficient revocation of users and attributes. This paper tackles the challenge of creating a secure, efficient, and verifiable access control scheme for fog-cloud computing that can handle multi-authority scenarios and dynamic revocation.
Methods
The proposed scheme, VO-MAACS, builds upon CP-ABE and incorporates several key innovations:
- Verifiable Outsourced Computation: The core idea is to offload the computationally intensive encryption and decryption processes to fog devices. Crucially, the paper introduces verification algorithms (
Verify_encandVerify_dec) that allow data owners and users to check if the fog devices have performed these computations correctly. This prevents "lazy" fog devices from returning incorrect results without detection. - Multi-Authority CP-ABE: The scheme supports multiple attribute authorities, which is essential for realistic fog-cloud environments where attributes can originate from various sources.
- Efficient Revocation: VO-MAACS addresses both user revocation and attribute revocation.
- User Revocation: This is handled efficiently by the Cloud Service Provider (CSP) simply deleting the revoked user's proxy key. This prevents the fog devices from performing outsourced decryption for that user without needing to re-encrypt existing data.
- Attribute Revocation: This is a more complex process involving two phases: key update and ciphertext re-encryption. The key update involves generating new attribute version keys and updating proxy keys and public attribute keys. The ciphertext re-encryption phase ensures that existing ciphertexts are updated to reflect the revoked attribute. The paper emphasizes that most of these update and re-encryption operations are outsourced to the CSP, minimizing the burden on data owners.
- System Model: The paper defines a three-level hierarchy: Cloud, Fog, and Devices. It also outlines a system model with six entities: Cloud Service Provider (CSP), Fog Devices (FDs), Global Certificate Authority (CA), Attribute Authorities (AAs), Data Owners (DOs), and Data Users (DUs).
Datasets
Not discussed in the paper. The performance evaluation uses simulated data based on access policies of varying complexity.
Key Insights and Ideas
- Decoupling Computation and Verification: The most significant insight is the separation of outsourced computation from its verification. This allows for the benefits of offloading to resource-constrained fog devices while maintaining security and trust through independent verification.
- Efficiency in Revocation: The proposed revocation mechanisms, particularly for users, are designed to be highly efficient. By simply removing a proxy key, the system effectively revokes access without costly re-encryption of all data. Attribute revocation, while more involved, also aims to minimize overhead by outsourcing most of the work.
- Fog as a Trusted Intermediate: The scheme positions fog devices as semi-trusted entities capable of performing computations but requiring verification. This aligns with the practical deployment of fog infrastructure.
- Addressing "Lazy" Servers: The explicit inclusion of verification algorithms directly tackles the "lazy server" problem, a common concern in outsourced computation scenarios.
Key Results
- Security: The scheme is claimed to be secure against collusion attacks and provides data confidentiality.
- Efficiency:
- Encryption/Decryption: The paper's simulations show that VO-MAACS requires significantly less time for user encryption and decryption compared to a baseline scheme (Lewko's scheme [14]). The outsourced encryption (
Encrypt_out) and decryption (Decrypt_out) times increase linearly with the number of attributes, but the user-side operations are substantially reduced.
- Encryption/Decryption: The paper's simulations show that VO-MAACS requires significantly less time for user encryption and decryption compared to a baseline scheme (Lewko's scheme [14]). The outsourced encryption (
- Verification: The verification of outsourced encryption (
Verify_enc) is shown to be efficient, with computation time increasing linearly with the number of attributes. - Revocation: The attribute revocation process shows reduced computing costs for Data Owners (DOs) compared to other schemes, as most of the overhead is outsourced to the CSP and AAs.
Limitations and Future Work
- Semi-trusted Fog Devices: While the scheme verifies computations, fog devices are still considered semi-trusted. They could potentially leak encrypted data, though they are assumed to execute assigned tasks.
- Performance with Large-Scale Systems: The simulations are conducted with a limited number of attributes and authorities. The performance in extremely large-scale fog-cloud environments with a vast number of devices and attributes might require further investigation.
- Future Work: The authors suggest that their scheme can be extended to support more complex access policies and potentially other cryptographic primitives. Further research into the security and performance in highly dynamic and heterogeneous fog-cloud environments would be beneficial.
Summary updated