Zamna: a tool for the secure and reliable storage, sharing, and usage of large data sets in data science applications

The exponential data growth (volume), continuously produced from different sources (variety) such as sensors, endusers, and IoT devices, has motivated the creation of efficient data science applications to obtain useful information (value and veracity) in critical decision-making processes (velocity). In health scenarios, the building of data science systems not only must include efficient modules for the acquisition, processing, and visualization of the data, but also transversal layers are required for the management, delivery and storage of data in a secure, reliable, and efficient manner in accordance with different norms or regulations for the handling of sensitive data. This is even more complex when the data science systems are distributed in environments of multiple infrastructures or in current and future computing models (for example, edge, fog, cloud or any combination of them). This article presents Zamna, a computational tool that allows end-users to support data science services by providing the resources for the secure and reliable data management, delivery and storage. Zamna manages the exchange and preparation of data in automatic and transparent manners, from acquisition until consumption by decision makers, which is a fundamental part in practically any data science application. Through its services, Zamna allows the automatic and transparent meeting of standards and regulations of data management, guaranteeing privacy, confidentiality, integrity, and availability of content, as well as tolerance to service failures and traceability. In a case study in the e-health domain, Zamna, allows achieving 70% compliance with international standards for the secure handling of sensitive data at the time that exhibited high performance due to the parallel patterns used in its construction.