Traceable Dynamic Public Auditing with Identity Privacy Preserving for Cloud Storage

Yinghui Zhang, Tiantian Zhang, Rui Guo, Shengmin Xu, Dong Zheng

Abstract

As organizations increasingly outsource data to cloud environments, the loss of physical control necessitates robust mechanisms for verifying data integrity. This paper presents a novel cryptographic protocol for "traceable dynamic public auditing" that balances the conflicting requirements of user privacy and accountability. By integrating threshold secret sharing with homomorphic linear authenticators, the scheme allows a third-party auditor to verify data integrity without compromising user anonymity. Crucially, the protocol decentralizes the power to deanonymize users, requiring a quorum rather than a single manager, while supporting efficient dynamic operations such as user joining and revocation.

Introduction: The Cloud Storage Trilemma

Outsourcing data to the cloud introduces significant security challenges. While "Public Auditing" allows a Third-Party Auditor (TPA) to verify data integrity without downloading the entire dataset, existing solutions often struggle with a "trilemma" of competing requirements:

  1. Identity Privacy: Users must remain anonymous to both the cloud provider and the TPA to prevent profiling or unauthorized tracking.
  2. Traceability: In cases of malicious data uploads, there must be a mechanism to identify the responsible party. However, giving a single "Group Manager" total power to deanonymize users creates a single point of failure and potential for administrative abuse.
  3. Dynamic Operations: Cloud groups are fluid, with users frequently joining or being revoked. Updating security keys for an entire group upon every membership change is often computationally prohibitive.

This paper addresses these challenges by proposing a decentralized trust model that ensures privacy, accountability, and scalability.

System Architecture

The proposed system model involves four primary entities: the Group Manager, the Users, the Cloud Storage Server, and the Third-Party Auditor (TPA). The interactions between these entities are designed to ensure that auditing can occur seamlessly while maintaining the security properties of the group.

Figure 1: System model for traceable dynamic public auditing in cloud storage.

As illustrated in Figure 1, the Group Manager handles system setup and user membership. Users interact with the Cloud Storage Server for data storage, while the TPA performs auditing tasks by issuing challenges and receiving proofs from the cloud. The system specifically accounts for dynamic group changes through "join" and "revocation" processes, ensuring that only current authorized members can generate valid authenticators for data blocks.

Methodology and Cryptographic Foundations

The protocol leverages several advanced cryptographic building blocks to achieve its goals:

  • Threshold Secret Sharing $(t, n)$: This is the core innovation for fair traceability. Instead of a single manager holding the de-anonymization key, the ability to trace a user is split among $n$ users. Identifying a signer requires a quorum of at least $t$ legitimate users to cooperate, preventing a malicious manager from unilaterally spying on group members.
  • Bilinear Pairings and BLS Signatures: These provide the mathematical foundation for short, verifiable signatures. These signatures serve as "authenticators"—metadata used by the TPA to check data blocks without accessing the raw data.
  • Polynomial Function Key Management: To handle dynamic user revocation, the group manager utilizes polynomial functions to update the group secret key. This allows non-revoked users to calculate new keys locally, while revoked users are mathematically excluded from the system without requiring a full system-wide re-keying.
  • Homomorphic Linear Authenticators (HLA): HLAs allow the TPA to aggregate multiple data block checks into a single, compact proof. This drastically reduces the communication overhead between the cloud and the auditor.

Performance Evaluation

The authors evaluated the scheme using a Java Pairing Based Cryptography (JPBC) library simulation, comparing it against existing frameworks (Huang’s and Hu’s schemes). Key findings include:

  • Signature Efficiency: Unlike previous schemes where signature generation time increased linearly with the number of users, this protocol maintains nearly constant signature time. At a group size of 150 users, the proposed scheme was approximately 10 times faster than Huang’s scheme.
  • Auditing Scalability: The computation cost for the TPA and the Cloud Server remains constant regardless of the group size. By avoiding expensive zero-knowledge proof calculations during the auditing phase, the system outperforms Hu’s scheme in high-volume data environments.
  • Communication Overhead: The total data exchanged during an auditing challenge is minimized, ensuring the protocol remains scalable for enterprise-level cloud storage.

Conclusion and Future Work

This research demonstrates that decentralizing trust through threshold cryptography can effectively resolve the tension between privacy and accountability in cloud auditing. By moving away from a single-manager de-anonymization model, the protocol provides a fairer environment for users.

Future work may address the logistical hurdles of the "quorum" requirement, ensuring that the necessary $t$ users are available for traceability when needed. Additionally, as quantum computing advances, transitioning these pairing-based features to lattice-based cryptography will be essential to maintain long-term security against post-quantum threats.

Summary updated

Abstract

KSII Transactions on Internet and Information Systems
BibTeX
@article{zhang2019traceable,
  author = {Yinghui Zhang and Tiantian Zhang and Rui Guo and Shengmin Xu and Dong Zheng},
  title = {{Traceable Dynamic Public Auditing with Identity Privacy Preserving for Cloud Storage}},
  journal = {Ksii Trans. Internet Inf. Syst},
  year = {2019},
  doi = {10.3837/tiis.2019.11.021},
}
Figure