Efficient Privacy-Preserving Access Control of Mobile Multimedia Data in Cloud Computing

Qi Li, Youliang Tian, Yinghui Zhang, Limin Shen, Jingjing GuoIEEE Access 2019 · DOI

Abstract

Ciphertext-policy attribute-based encryption (CP-ABE) is a powerful tool for cloud-assisted data sharing, but it often compromises user privacy by exposing access policies in plaintext and imposes heavy computational burdens on mobile devices. This paper presents PPCMM, a privacy-preserving scheme that utilizes partially hidden access policies and an online/offline encryption split. By offloading complex decryption tasks to the cloud, PPCMM achieves constant-time user-side decryption and high efficiency, all while maintaining adaptive security in the standard model.

1. Introduction

The proliferation of mobile multimedia sharing via cloud services has heightened the need for robust access control. While Ciphertext-Policy Attribute-Based Encryption (CP-ABE) allows data owners to specify access rules, it faces two primary hurdles: the "privacy paradox" and computational intensity. Standard CP-ABE reveals access policies (e.g., "Department: Cardiology") in plaintext, allowing unauthorized parties to infer sensitive information about the data or the intended recipient. Furthermore, the bilinear pairings and exponentiations required by CP-ABE are taxing for battery-powered mobile devices, often leading to significant battery drain and slow performance.

2. The PPCMM System Architecture

The proposed PPCMM (Privacy-Preserving Cloud-assisted Mobile Multimedia) framework addresses these challenges through a collaborative architecture involving the Mobile Data Owner (MDO), an Attribute Authority (AA), a Cloud Service Provider (CSP), and the Mobile Data User (MDU).

The PPCMM system architecture illustrating the interaction between the mobile data owner (MDO), cloud service provider (CSP), and mobile data user (MDU).

In this model, the MDO prepares data for sharing, the AA manages cryptographic keys, and the CSP acts as a semi-trusted intermediary that stores data and performs pre-decryption tasks to save the MDU's resources. This structure allows the system to handle complex multimedia files without overwhelming the mobile hardware.

3. Privacy-Preserving Mechanisms

A core innovation of PPCMM is the partially hidden access policy. Unlike traditional schemes that expose full attributes, PPCMM splits attributes into names and values. This approach is based on Linear Secret Sharing Schemes (LSSS) for flexible logic.

A partially hidden access policy where attribute names are visible but sensitive values are masked.

As illustrated in the policy tree above, attribute names (e.g., "Department") are left in plaintext to facilitate cloud-side matching, while the sensitive values (e.g., "Cardiology") are hidden within the ciphertext. This ensures that even if the CSP or an attacker intercepts the policy, they cannot determine the specific sensitive traits required for access, effectively preserving user anonymity.

4. Computational Efficiency: Online/Offline Split

To accommodate resource-constrained mobile hardware, PPCMM bifurcates the encryption process into two distinct phases:

  • Offline Phase: When the device is idle or charging, it pre-calculates intermediate ciphertext components that are independent of the specific message or access policy.
  • Online Phase: When a user initiates a share, the device performs only lightweight arithmetic to bind the data to the policy, resulting in near-instantaneous encryption regardless of policy complexity.

Similarly, decryption is optimized through outsourcing. The CSP performs a "Matching Test" to verify if the MDU's attributes satisfy the hidden policy. If successful, the CSP generates a partial decryption. The MDU then only needs to perform a single modular exponentiation to recover the final plaintext, ensuring that user-side decryption time remains constant.

5. Performance and Security Results

The authors implemented PPCMM using the Java Pairing Based Cryptography (JPBC) library and benchmarked it against existing state-of-the-art schemes, specifically the ZZD and ZHMR schemes. Experimental results demonstrated that while competing schemes showed linear growth in decryption time as the number of attributes increased (exceeding 600ms for 100 attributes), PPCMM’s user-side decryption remained flat and negligible.

Security-wise, PPCMM is proven adaptively secure in the standard model. This provides a higher level of security than the "selective" models used in many previous works, ensuring protection even when attackers can choose their targets after observing system parameters.

6. Limitations and Future Work

While PPCMM significantly improves efficiency and privacy, it assumes an "honest-but-curious" cloud provider. Future iterations aim to incorporate verifiability, allowing users to cryptographically prove that the cloud performed the outsourced decryption correctly. Additionally, the authors plan to explore more efficient methods for attribute revocation and investigate the impact of malicious cloud behavior on data integrity.

Summary updated

Abstract

With the nature of allowing the encryptor to define the access policy before encrypting a message, Ciphertext-policy attribute-based encryption (CP-ABE) has been widely adopted as a primitive to design cloud-assisted mobile multimedia data sharing system. However, in most previous works, the access policy sent along with ciphertext remains in the palintext form, which would expose the user's privacy to anyone who can get the ciphertext even if he is not authorized to decrypt. In addition, the resource-constrained mobile devices can not carry out frequent encryption and decryption task caused by CP-ABE. Such drawbacks may reduce the enthusiasm of consumers to share the multimedia data via their mobile devices. In this paper, we proposed PPCMM, a privacy-preserving cloud-assisted mobile multimedia data sharing scheme, where each attribute is described by an attribute name and an attribute value. The attribute values are embedded in the ciphertext and only the attribute names are revealed in the access policy. The encryption is divided to two phase: online and offline. In the offline stage, the data owner can prepare the intermediate ciphertext components. Once receiving the encryption requirement of a specific access policy and the multimedia data, the data owner can quickly form the final legal ciphertext in the online stage. By employing the decryption outsourcing technique, most computation overhead of matching test and decryption is offload to the cloud server. The security proof showed that PPCMM is adaptively secure in the standard model. The performance analysis indicated that PPCMM greatly reduces the computation cost in both online encryption and user decryption.
BibTeX
@inproceedings{li2019efficient,
  author = {Qi Li and Youliang Tian and Yinghui Zhang and Limin Shen and Jingjing Guo},
  title = {{Efficient Privacy-Preserving Access Control of Mobile Multimedia Data in Cloud Computing}},
  booktitle = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2939299},
}
Figure