Design for Prevention of Intranet Information Leakage via Emails

The ubiquitous presences of internet and network technologies have enabled electronic mail systems as the primary medium of communication. Both between and within organizations, sensitive and personal information often transits through the electronic mail systems undetected. Information leakage through this mode of communication has become a daunting problem in today's world. Often the mail volume within an organization is quite large making manual monitoring impossible. In this paper an integration of secure information flow techniques on intranet electronic mail systems is investigated. Categorization of emails based on the sensitivity is accomplished effectively using machine learning techniques. Analyzing the information flow and simultaneously mapping, categorizing and sorting emails in real time prior to receipt of emails has been characterized in this study. Defining security policies and application of lattice models for controlled exchange of emails is discussed. The paper proposed a secure architecture for an email web application. Experimental analysis on the accuracy of the application was determined using Enron email dataset.