Toward Scalable Anonymization for Privacy-Preserving Big Data Publishing

Big data is collected and processed using different sources and tools, which leads to privacy issues. Privacy-preserving data publishing techniques such as k-anonymity, l-diversity, t-closeness are used to de-identify data, but chances of re-identification are there as data is collected from multiple sources. Due to a large amount of data, less generalization or suppression is required to achieve same level of privacy, which is also known as "large crowd effect," but to handle such a large data for anonymization is also a challenging task. MapReduce handles a large amount of data, but it distributes data into small chunks, so the advantage of large data cannot be achieved. Therefore, scalability of privacy-preserving techniques has become a challenging area of research, and we are trying to explore it by proposing an algorithm for scalable k-anonymity for MapReduce. Based on comparison with existing algorithm, our approach shows significant improvement in running time.