Revisit: Multi-Tree Approach to Mutable Order-Preserving Encoding

Abstract

In order-preserving encryption, it was discovered that utilising a hypergeometric probability distribution exposed information about half of the plaintext bits and their respective distances. To address this vulnerability, mutable order-preserving encoding was introduced as a solution, but it still suffered from revealing the plaintext distribution and exhibiting high correlation. In an effort to overcome these limitations, Lee and Jho proposed a multi-tree approach that utilises the binary tree invariant and its inverse, resulting in increased complexity for client operations. In this paper, we propose a novel multi-tree approach to mutable order-preserving encoding that, by introducing a secret offset at the client side, not only simplifies the server-side structure but also reduces overall system complexity. To validate the effectiveness of our approach, we conducted comprehensive experiments using the TPC-C benchmark dataset, evaluating encoding time and query performance across varying data scales. The results demonstrate that our scheme achieves 36% higher encoding efficiency compared to existing multi-tree methods.