Privacy-Preserving Sorting Algorithms Based on Logistic Map for Clouds

Hua Dai,Zhiye Chen,Xun Yi,Geng Yang,Hui Ren

doi:10.1155/2018/2373545

Abstract

Outsourcing data in clouds is adopted by more and more companies and individuals due to the profits from data sharing and parallel, elastic, and on-demand computing. However, it forces data owners to lose control of their own data, which causes privacy-preserving problems on sensitive data. Sorting is a common operation in many areas, such as machine learning, service recommendation, and data query. It is a challenge to implement privacy-preserving sorting over encrypted data without leaking privacy of sensitive data. In this paper, we propose privacy-preserving sorting algorithms which are on the basis of the logistic map. Secure comparable codes are constructed by logistic map functions, which can be utilized to compare the corresponding encrypted data items even without knowing their plaintext values. Data owners firstly encrypt their data and generate the corresponding comparable codes and then outsource them to clouds. Cloud servers are capable of sorting the outsourced encrypted data in accordance with their corresponding comparable codes by the proposed privacy-preserving sorting algorithms. Security analysis and experimental results show that the proposed algorithms can protect data privacy, while providing efficient sorting on encrypted data.

Highlights

With the profits from data sharing and parallel, elastic, and on-demand computing, clouds are becoming more and more popular with companies and individuals
We use the same datasets to evaluate the privacy-preserving sorting algorithms based on the logistic map and orderpreserved encryption (OPE)
The reason is that the output data of OPE, which is used for privacy-preserving sorting, is more complex than the comparable codes generated by the logistic map

Summary

Introduction

With the profits from data sharing and parallel, elastic, and on-demand computing, clouds are becoming more and more popular with companies and individuals. As one of the most important technologies, machine learning is very useful and wildly adopted in many areas, such as prediction [1, 2] and multimedia data processing [3, 4]. It usually utilizes huge data volume, such as wireless multimedia data and human health data, to build intelligent models and systems for practical applications. Due to the need of large and elastic scale of storage and computing resources, those huge volume data are usually processed in clouds [5,6,7]. Data owner (DO) outsources their data in the cloud server (CS) for on-demand services which enhance the efficiency of complex computation such as machine learning and save the hardware/software cost

Methods

Results

Conclusion