Decentralizing Privacy: Using Blockchain to Protect Personal Data

Guy Zyskind, Oz Nathan, Alex PentlandIEEE Symposium on Security and Privacy Workshops 2015 · DOI

Abstract

In the modern digital economy, personal data is often treated as currency, stored in centralized silos where users lack control and transparency. This paper proposes a decentralized system that shifts ownership back to the user. By utilizing a blockchain as an automated, trustless access-control manager and a Distributed Hash Table (DHT) for encrypted off-chain storage, the authors create a platform where users can grant, monitor, and revoke data access in real-time without relying on a central authority.

Introduction

Current digital services often require users to trade their personal data for access. This data is typically stored in centralized silos (e.g., Facebook or Google), leading to three core issues: a lack of true ownership, a lack of transparency regarding data usage, and coarse, "all-or-nothing" access controls. While privacy-preserving techniques like k-anonymity exist, they are often vulnerable to de-anonymization or are computationally prohibitive. This research addresses these gaps by decoupling data storage from access control logic.

System Architecture

The authors introduce a hybrid architecture that leverages the unique strengths of blockchain and decentralized storage.

  • Blockchain as a Policy Engine: Instead of storing heavy data on-chain, the blockchain acts as a trustless auditor. It manages access control policies and stores "pointers" (hashes) to the data. Two primary transaction types are introduced: $T_{access}$ for managing permissions and $T_{data}$ for data retrieval.
  • Off-Chain Storage (DHT): Encrypted data is stored in a Distributed Hash Table (DHT). Only the user and authorized services possess the keys to decrypt this information.
  • Compound Identities: To enhance pseudonymity, the system generates unique identities for every user-service pair. This ensures that data generated for one application cannot be easily linked to another.
System architecture showing the interaction between users, services, the blockchain policy engine, and the DHT storage.

When a service requests data, the blockchain nodes verify the digital signature against the stored policy. If authorized, the service receives a pointer to the encrypted data in the DHT.

Key Insights

  • Trustless Auditing: The blockchain creates a permanent, public, yet pseudonymous record of every data access event, providing unprecedented transparency.
  • Decoupled Storage: By keeping data off-chain, the system avoids the scalability and latency issues inherent in storing large datasets directly on a ledger.
  • User-Centric Revocation: Unlike traditional mobile permissions, this system allows for immediate revocation. Because services must query the blockchain for a pointer every time they access data, a policy change in a $T_{access}$ transaction immediately "locks" the data from further use.

Security and Implementation

The paper provides a technical framework demonstrating secure identity management and access control. The security analysis posits that as long as the underlying blockchain remains tamper-free, adversaries cannot impersonate users or extract information from the public ledger, as it contains only hashed pointers.

Limitations and Future Work

While the system secures storage and retrieval, a service could potentially store a local copy of raw data once queried. To mitigate this, the authors suggest integrating Secure Multi-party Computation (MPC). This would allow services to run computations on data shares across a network (NET) without ever seeing the raw data itself.

Proposed Secure Multi-party Computation (MPC) protocol for decentralized data processing.

Future work also includes addressing the latency of blockchain transactions for real-time applications and developing trust-based weighting systems for DHT nodes to prevent malicious data deletion or Sybil attacks.

Summary updated

BibTeX
@inproceedings{zyskind2015decentralizing,
  author = {Guy Zyskind and Oz Nathan and Alex Pentland},
  title = {{Decentralizing Privacy: Using Blockchain to Protect Personal Data}},
  booktitle = {IEEE Symposium on Security and Privacy Workshops},
  year = {2015},
  doi = {10.1109/spw.2015.27},
}
Figure