A policy-based containerized filter for secure information sharing in organizational environments

In organizational environments, sensitive information is unintentionally exposed and sent to the cloud without encryption by insiders that even were previously informed about cloud risks. To mitigate the effects of this information privacy paradox, we propose the design, development and implementation of SecFilter, a security filter that enables organizations to implement security policies for information sharing. SecFilter automatically performs the following tasks: (a) intercepts files before sending them to the cloud; (b) searches for sensitive criteria in the context and content of the intercepted files by using mining techniques; (c) calculates the risk level for each identified criterion; (d) assigns a security level to each file based on the detected risk in its content and context: and (e) encrypts each file by using a multi-level security engine, based on digital envelopes from symmetric encryption, attribute-based encryption and digital signatures to guarantee the security services of confidentiality, integrity and authentication on each file at the same time that access control mechanisms are enforced before sending the secured file versions to cloud storage. A prototype of SecFilter was implemented for a real-world file sharing application that has been deployed on a private cloud. Fine-tuning of SecFilter components is described and a case study has been conducted based on document sharing of a well-known repository (MedLine corpus). The experimental evaluation revealed the feasibility and efficiency of applying a security filter to share information in organizational environments. (C) 2019 Elsevier B.V. All rights reserved.