Privacy preserving and performance analysis on not only SQL database aggregation in bigdata era
Dong‐Her Shih, Feng-Chuan Huang, Wei-Hao Lai, Ming‐Hung Shih · 2017
Database management systems have been indispensable to enterprises for decades. As the amount of data dramatically increased, database aggregation has encountered a dilemma between privacy and performance. In traditional database aggregation, all attributes have been encrypted to protect the privacy of data. However, in big data, this privacy measure is no longer feasible because cryptography will degrade the system performance. Therefore, a database aggregation scheme that not only protects privacy but also delivers high performance is in need. In this study, we propose a partial encryption method that identifies the most important privacy attributes during database aggregation in order to provide both privacy protection and good performance. In addition, we study the performance differences of our method when deployed on both traditional SQL databases and Not Only SQL (NoSQL) databases (HBase and Cassandra), by using Yahoo! Cloud Service Benchmark (YCSB) for performance evaluation. Our experiments provided guidance for future developers to consider the tradeoff between privacy and performance in different database systems for big data analytics.