Authors:
Mohammad Anagreh
1
;
2
;
Peeter Laud
1
and
Eero Vainikko
2
Affiliations:
1
Cybernetica, Mäealuse 2/1, 12618 Tallinn, Estonia
;
2
Institute of Computer Science, University of Tartu, Narva maantee 18, 51009 Tartu, Estonia
Keyword(s):
Privacy-preserving Computation, Secure Multi-party Computation (SMC), Shortest Path Problem, Single- Instruction-Multiple-Data (SIMD), Breadth-first Search, Sharemind.
Abstract:
Reducing the round complexity in secure multiparty computation (SMC) protocols is a worthy goal due to the latency of the network. The SIMD approach is considered an efficient strategy to reduce the round complexity of an SMC protocol. This paper studies the secure multiparty computation (SMC) protocol for the shortest path problem in sparse and dense graphs, building upon the breadth-first search algorithm. The sensitivity of operations in processing the algorithms led us to produce two different structural algorithms for computing the shortest path. We present state-of-the-art parallel privacy-preserving shortest path algorithms for weighted and unweighted graphs based on the breadth-first search. We have implemented the proposed algorithms on top of the Sharemind SMC protocol set and tested it for different graphs, dense and sparse, represented as the adjacency matrix.