Parallel Privacy-Preserving Shortest Paths by Radius-Stepping (original) (raw)

2021 29th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP), 2021

Abstract

The radius-stepping algorithm is an efficient, parallelizable algorithm for finding the shortest paths in graphs. It solved the problem in triangle\triangletriangle-Stepping algorithm, which has no known theoretical bounds for general graphs. In this paper, we describe a parallel privacy-preserving method for finding Single-Source Shortest Paths (SSSP). Our optimized method is based on the Radius-Stepping algorithm. The method is implemented on top of the Secure Multiparty Computation (SMC) Sharemind platform. We have re-shaped the radius-stepping algorithm to work on vectors representing the graph in a SIMD manner, in order to enable a fast execution using the secret-sharing based SMC protocol set of Sharemind. The results of the real implementation show an efficient method that reduced the execution time hundreds of times in comparison with a standard case of the privacy-preserving radius-stepping and triangle\triangletriangle-Stepping algorithms.

Eero Vainikko hasn't uploaded this paper.

Let Eero know you want this paper to be uploaded.

Ask for this paper to be uploaded.