Record Level Splitting of Anonymized Tabular Data on to Multicloud for Privacy Preservation

Abstract

Tabular data is the principal form of high-dimensional data in the cloud storage, and it contains sensitive attributes. Privacy Preservation (PP) is a crucial task in the storage and provision of such data (like health data) for different research and analysis purposes. Data encryption provides sturdy security, but the demand for high computation power limits it not to support the focused High-dimensional Tabular Data (HTD). Data Splitting through Anonymization (DSA) or packetization is the low computation and reasonable PP alternative in use on HTD. However DSA is vulnerable to the major attacks like attribute disclosure, similarity attack, sensitivity attack, etc., this study shows that incorporating record level splitting on to multiple clouds makes DSA a resilient and low computation PP method. The privacy preservation will include a mostly single attribute in majority cases and anonymization will be applied to it. The situation like combining data that got segregated into multiple clouds is tough and dealing with multiple sensitive attributes is another major thing that needs to be studied and dealing with the data that update’s constantly can also be used in this which mainly used in hospital data and some company data that update data at some interval of time.