Abstract
Data and data processing have become an indispensable aspect for our society. Insights drawn from collective data make invaluable contribution to scientific and societal research and business. But there are increasing worries about privacy issues and data misuse. This has prompted the emergence of decentralised personal data stores (PDS) like Solid that provide individuals more control over their personal data. However, existing PDS frameworks face challenges in ensuring data privacy when performing collective computations with data from multiple users. While Secure Multi-Party Computation (MPC) offers input secrecy protection during the computation without relying on any single party, issues emerge when directly applying MPC in the context of PDS, particularly due to key factors like autonomy and decentralisation. In this work, we discuss the essence of this issue, identify a potential solution, and introduce a modular architecture, Libertas, to integrate MPC with PDS like Solid, without requiring protocol-level changes. We introduce a paradigm shift from an 'omniscient' view to individual-based, user-centric view of trust and security, and discuss the threat model of Libertas. Two realistic use cases for collaborative data processing are used for evaluation, both for technical feasibility and empirical benchmark, highlighting its effectiveness in empowering gig workers and generating differentially private synthetic data. The results of our experiments underscore Libertas' linear scalability and provide valuable insights into compute optimisations, thereby advancing the state-of-the-art in privacy-preserving data processing practices. By offering practical solutions for maintaining both individual autonomy and privacy in collaborative data processing environments, Libertas contributes significantly to the ongoing discourse on privacy protection in data-driven decision-making contexts.
Introduction
The digital economy thrives on data, but the centralization of personal information creates significant privacy risks. Decentralized Personal Data Stores (PDS), such as the Solid project, have emerged to return control to individuals by allowing them to store data in personal "Pods." However, a critical challenge remains: how can we perform collective computations—such as calculating the average salary across a community—without requiring users to surrender their data to a central aggregator?
While Secure Multi-Party Computation (MPC) provides a cryptographic path forward, its traditional application often assumes an "omniscient" view where a single entity dictates security parameters and trusted servers. This contradicts the decentralized ethos of PDS. To address this, Libertas identifies five fundamental requirements for privacy-preserving collective computation in decentralized environments.
The Libertas Architecture
Libertas is a modular architecture designed to bridge PDS and MPC without requiring protocol-level changes. Its core innovation is the Delegated-Decentralised MPC model, which shifts the trust paradigm from platform-centric to user-centric.
In this model, users define their trust preferences within their Pods, specifying which Encryption Agents (EAs) and Computation Agents (CAs) they trust. EAs act as trusted intermediaries that fetch raw data and convert it into cryptographic "secret shares." These shares are then processed by CAs, who perform the computation without ever seeing the original data. This separation of duties allows for a flexible, reusable infrastructure where different organizations can host different agents based on user trust.
Evaluation and Scalability
The system was evaluated using realistic scenarios, including analyzing gig worker income and generating differentially private synthetic data via the MWEM algorithm. A key finding was Libertas' linear scalability ($O(n)$). Unlike "direct" MPC models where complexity grows polynomially with the number of participants, the delegated model handled over 500 clients efficiently. In the MWEM use case, the system successfully processed data from 1,000 providers, demonstrating its feasibility for large-scale applications.
However, benchmarks also revealed that as the number of users grows, the "setup time" for establishing secure connections becomes a primary bottleneck, suggesting a clear direction for future optimization.
Conclusion and Future Work
Libertas demonstrates that it is possible to maintain individual autonomy and privacy while enabling powerful collective analytics. By allowing users to express heterogeneous trust preferences, it aligns cryptographic security with the decentralized nature of the modern web. Future research will focus on developing automated trust discovery mechanisms, evaluating performance across high-latency Wide Area Networks (WAN), and addressing the governance challenges inherent in decentralized modular architectures.
Summary updated