一种最小化安全多方计算任务的方法

A method to optimize the efficiency of secure multiparty computation

摘要:

安全多方计算（Secure multi-party computation：MPC）允许在不公开各参与方私有数据的情况下完成联合计算。然而，现有的计算任务往往涉及到多方海量数据集的分析与处理，使得MPC的实际可用性显著降低。提高MPC数据处理体量，是目前研究的主要方向之一。为提高MPC处理大规模数据的能力，将MPC算法与数据并行分析框架相结合，基于最小化多方计算任务的思想，提出安全多方计算效率优化技术。创建算法的有向无环图，标注MPC节点及非MPC节点，采用静态分析、查询重写转换和分区启发式等技术，最小化MPC计算量，提高计算的并发程度。以多方线性回归为例，讨论适应大数据分析的安全多方计算技术。实验结果表明提出的安全多方计算优化技术在确保计算精度的条件下能够显著降低计算耗时。算法提高了系统的效率，增强了MPC的实用能力。

Abstract:

Secure multi-party computing (MPC) allows joint computing without disclosing the private data of each participant. However, the existing computing tasks often involve the analysis and processing of multi-party massive data sets, which significantly reduces the actual availability of MPC. Improving the volume of MPC data processing is one of the main research directions at present. In order to improve the ability of MPC to process large-scale data, the MPC algorithm is combined with the data parallel analysis framework. Based on the idea of minimizing multi-party computing tasks, a secure multi-party computing efficiency optimization technology is proposed. The directed acyclic graph of the algorithm is created, the MPC nodes and non MPC nodes are marked, and the techniques of static analysis, query rewriting transformation and partition heuristic are used to minimize the amount of MPC calculation and improve the concurrency of calculation. Taking multi-party linear regression as an example, this paper discusses the secure multi-party computing technology suitable for big data analysis. The experimental results show that the proposed secure multi-party computing optimization technology can significantly reduce the computing time under the condition of ensuring the computing accuracy. The algorithm improves the efficiency of the system and enhances the practical ability of MPC.