Combining Fragmentation and Encryption to Ensure Big Data at Rest Security Houyem Heni 1( ✉ ) , Marwa Ben Abdallah 2 , and Faiez Gargouri 1 1 MIR@CL Laboratory, University of Sfax, Sfax, Tunisia houyem.heni@gmail.com, faiez.gargouri@isimsf.rnu.tn 2 Habib Cheikhrouhou, Ariana, Tunisia marouabenabdalla@gmail.com Abstract. We are in the midst of a revolution within computing, it goes under the name of big data. Thus, Due to big data proliferation and the various infor‐ mation resources, our personal data will be shared and published by all people; that is why our privacy will be increasingly accessed, and thus threatened by hackers. In this context, many researchers have proposed different methods to ensure the security of sensitive and identifiable information. Through this paper, we want to dig into the security context while implementing a methodological approach to protect the sensitive data in the big data frameworks. In this article, we propose a method which combines fragmentation and encryption to ensure security in Mongo database. It allows sensitive data security in NoSQL context. Keywords: Big data · NoSQL · Security · Mongo db · Data fragmentation 1 Introduction With the explosion of the amounts of digital data in various fields known as Big Data, the need to mention data security importance is felt. Security is important in computer science research, particularly for data bases which store critical, personal and identifiable information such as credit card number, medical records, and social number. They are used to represent a variety of data about a domain as a set of concepts, using a shared vocabulary to point out the types and properties of these concepts and their relationships. NoSQL databases are in the base of storage of Big Data. They permit to store large volumes of structured, semi-structured, and unstructured data. These databases suffer from lack of robust methods of security since they are able to handle unstructured data. To resolve this issue, we propose to split sensitive data and confidential collections of NoSQL database and save these splits separately [1]. In this paper, we consider that a methodological approach for big data security based on data splitting and collections fragmentation is highly needed. The main motivation of this work is the ability to reason on characteristics of the security solutions and to retrieve data in an efficient and distributed manner. We propose an approach where the target mechanism is represented as a data splitting and the source corresponds to a © Springer International Publishing AG, part of Springer Nature 2018 A. Abraham et al. (Eds.): HIS 2017, AISC 734, pp. 177–185, 2018. https://doi.org/10.1007/978-3-319-76351-4_18