BUAA-BDA/OpenHuFu

What it solves

OpenHuFu is designed to address the problem of data isolation. It allows for secure query processing and analytics over a federation of multiple data owners without requiring the raw data to be shared, which is often prohibited by security and privacy concerns.

How it works

The system implements a data federation architecture where queries are processed across multiple owners. It leverages secure multi-party computation (SMC) techniques—including secret sharing, garbled circuits, and oblivious transfer—to ensure that data remains private while still allowing for complex queries. The system is extensible, allowing researchers to implement their own custom aggregation and join algorithms.

Who it’s for

It is primarily intended for researchers who need a framework to quickly implement, test, and evaluate the performance of new federated query processing algorithms using SMC techniques over benchmark datasets.

Highlights

• Broad Query Support: Supports a wide range of relational queries (filter, projection, join, aggregate, sort, limit) and specialized spatial queries (range, KNN, distance join).
• SMC Integration: Built-in support for secret sharing, garbled circuits, and oblivious transfer.
• Extensibility: Provides Java interfaces for developers to easily add new OwnerAggregateFunction and OwnerJoin implementations.
• Performance Monitoring: Includes tools to evaluate communication costs and running time (local, encryption, and decryption).
• Data Generation: Includes scripts for generating both TPC-H relational data and synthetic spatial data.