Fast and Secure Multiparty Querying over Federated Graph Databases

Area of Science:

• Computer Science
• Cybersecurity
• Database Systems

Background:

• Federated graph databases present unique challenges for privacy-preserving data analysis.
• Existing solutions for multi-party querying often compromise on efficiency or security.

Purpose of the Study:

• To develop an efficient framework for privacy-preserving multi-party querying (PPMQ) over federated graph databases.
• To enhance data security through Secure Multi-Party Computation (SMPC) protocols.

Main Methods:

• Developed a PPMQ framework with two distinct security protocols: client-based and server-based.
• The server-based protocol integrates encrypted hashing for augmented security.
• Employed an honest but curious security model.

Main Results:

• PPMQ demonstrates execution times and overheads comparable to Neo4j Fabric.
• PPMQ significantly outperforms previous systems like SMPQ and Conclave in efficiency.
• The enhanced server protocol offers improved robustness against brute force attacks.

Conclusions:

• PPMQ provides a superior solution for privacy-preserving multi-party querying in federated graph databases.
• The framework achieves a strong balance between computational efficiency and robust data privacy.
• PPMQ enhances security guarantees beyond existing methods.