Distributed Differential Evolution for Anonymity-Driven Vertical Fragmentation in Outsourced Data Storage

Abstract

Vertical fragmentation is a promising technique for outsourced data storage. It can protect data privacy while conserving original data without any transformation. Previous vertical fragmentation approaches need to predefine sensitive associations in data as the optimization objective, therefore unavailable for the data lacking related prior knowledge. Inspired by the anonymity measurement in anonymity approaches such as k-anonymity, an anonymity-driven vertical fragmentation problem is defined in this paper. To tackle this problem, a set-based distributed differential evolution (S-DDE) algorithm is proposed. An island model containing four sub-populations is adopted to improve population diversity and search efficiency. Two set-based update operators, i.e., set-based mutation operator and set-based crossover operator, are designed to transfer the calculation of discrete values to corresponding sets in vertical fragmentation. Extensive experiments are carried out, and the performance of S-DDE on anonymity-driven vertical fragmentation is verified. The computation efficiency of S-DDE is investigated, and the effectiveness of the generated vertical fragmentation solution by S-DDE is confirmed.