DATA ANALYSIS AND DESIGN

Data Analysis and Design



Data Analysis and Design

Introduction

Information storage and manipulation using databases have become crucial to the day-today activities of most enterprises. Although databases have many advantages, protecting the data contained in them is a difficult task. One reads more and more about security leaks. Some solutions have been proposed to incorporate security into database management systems. But these solutions are not sufficient to satisfy the stringent requirements of military applications (Lampson, 2002, pp. 18). Multilevel security is being recommended for such applications. In this paper, we will discuss the notion of a Multilevel Secure Database Management System (MLS/DBMS) and then describe the difficulties in designing such a system. We will also propose a design for an MLS/DBMS. The model that we have used to represent the database is the relational model; finally, we will describe how multilevel security concepts developed for DBMSs can be applied to knowledge base management systems (KBMSs).

Multilevel secure database management systems

In an MLS/DBMS, it is possible that not all of the data contained in the database is equally sensitive. However, present-day DBMSs are not built with adequate controls and mechanisms to assure that users are allowed to access only the data for which they have been granted a clearance. Thus, an MLS/DBMS is different from a conventional DBMS in at least the following two ways:

Every data item in the database has associated with it a security level,

A user's access to data must be controlled based upon the user's clearance.

Providing an MLS/DBMS service on current computing systems presents many problems. It is only recently and with considerable difficulty that multilevel security has been incorporated into operating systems. This solution is called a Trusted Computing Base (TCB). Although hosting an MLS/DBMS on a TCB overcomes many of the compromises that could be made, there are still many problems that need to be resolved. The most obvious of these problems is that the granularity of classification in a DBMS is generally finer than a file and may be as fine as a single data element in a file. Another problem that is unique to databases is the necessity to classify data based on content, time, aggregation and context. Furthermore, DBMSs are also vulnerable to subtle covert channel attacks in which Trojan horses within the DBMS encode sensitive information in benign fields, and also to inference attacks where a user infers unauthorized data from the knowledge that he has accumulated (Graham, 1972, pp. 29).

Database Security

Security policy

The issues addressed by a security policy for an MLS/DBMS should include mandatory access, classification by context, content, and time, functional manipulations, aggregation, inference, downgrading, discretionary access and integrity. Each of these issues will be briefly described below. Mandatory access control requires that users only access the data to which they possess a clearance. Content-based classification assigns security levels to the data depending on their content. Context-based classifications assign security levels to the data depending on the particular context in which the data is ...