Scalable Distributed Data Anonymization

TL;DR: In this article, the authors present an approach for enabling a distributed anonymization process over large collections of sensor data using an arbitrary number of workers within the Spark framework, which is scalable and does not affect the quality of the anonymized dataset.

Abstract: We present an approach for enabling a distributed anonymization process over large collections of sensor data. Our approach anonymizes large datasets (which might not fit in main memory) using an arbitrary number of workers within the Spark framework. We describe how to parallelize the anonymization process through a proper partitioning of the dataset. Our experimental evaluation shows that the proposed approach is scalable and do not affect the quality of the anonymized dataset.