Abstract
The recent years have seen the birth of several NoSQL datastores, which are getting more and more popularity for their ability to handle high volumes of heterogeneous and unstructured data in a very efficient way. In several cases, NoSQL databases proved to outclass in terms of performance, scalability, and ease of use relational database management systems, meeting the requirements of a variety of today ICT applications. However, recent surveys reveal that, despite their undoubted popularity, NoSQL datastores suffer from some weaknesses, among which the lack of effective support for data protection appears among the most serious ones. Proper data protection mechanisms are therefore required to fill this void. In this work, we start to address this issue by focusing on access control and discussing the definition of a fine-grained access control framework for document-oriented NoSQL datastores. More precisely, we first focus on issues and challenges related to the definition of such a framework, considering theoretical, implementation, and integration aspects. Then, we discuss the reasons for which state-of-the-art fine-grained access control solutions proposed for relational database management systems cannot be used within the NoSQL scenario. We then introduce possible strategies to address the identified issues, which are at the basis of the framework development. Finally, we shortly report the outcome of an experience where the proposed framework has been used to enhance the data protection features of a popular NoSQL database.
Highlights
NoSQL datastores are getting popularity in a variety of scenarios, and their diffusion is growing especially within the data management back-end of modern web applications, and the data storage and analysis layer of Internet of Things platforms
We focus on access control features of NoSQL datastores, since access control is the core data protection module of any DBMS
The analysis described in this paper is partially based on early research experiences on NoSQL datastores that we did with MongoDB [10, 12], as well as on ongoing research activities finalized to the generalization of the approach in [10, 12]
Summary
NoSQL datastores are getting popularity in a variety of scenarios, and their diffusion is growing especially within the data management back-end of modern web applications, and the data storage and analysis layer of Internet of Things platforms. Recent surveys on database popularity rank documentoriented NoSQL datastores in the top position This is probably due to the flexibility of these systems, the provided advanced analysis features, and the native support for the management of JSON8 data, which, at present, is among the most common data exchange format of modern applications. For these reasons, in this work, we target document-oriented datastores. As we will discuss throughout the paper, the schemaless data model of document-oriented datastores do not allow to straightforwardly reuse the FGAC enforcement mechanisms defined for RDBMSs. so far no standard NoSQL query language has emerged yet (neither in general nor for a specific datastore category), and each datastore adopts a different language.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
