| The explosive growth of the Internet and network connections has led to increased opportunities for attackers to access computer systems illegitimately. As with any real world processes, almost every electronic commerce site uses a Database Management System (DBMS) to store and provide access to information. With the increase in attacks, the need to provide security for these DBMS systems has become critical. While several techniques for providing security have been proposed, very limited research has been done on how to survive a set of successful database attacks that can seriously impair the integrity and availability of a database. Most traditional secure systems cannot adjust their behavior after deployment. As a result, they perform very poorly under certain environmental changes, and their trustworthiness can dramatically change from time to time. Therefore, it is important for survivable systems to have the ability to adapt themselves to their environment.;Intrusion Tolerant Database System (ITDB) is a new paradigm for secure database systems that can detect intrusions, isolate attacks, contain damage, and access and repair damage caused by intrusions. What makes ITDB superior to conventional secure approaches is that it has an ability to reconfigure. Thus, it can yield much more stabilized levels of trustworthiness under environmental changes. However, the reconfiguration faces the problem of finding the best system configuration because it must choose out of a very large number of configuration sets and under multiple conflicting criteria, which is a NP-hard (Non-deterministic Polynomial-time hard) problem.;This study focuses on two aspects of addressing adaptation problems in ITDB. First, a rule-based mechanism and neuro-fuzzy technique are proposed to apply to the adaptation model so that ITDB can intelligently reconfigure itself to changing environments and efficiently resolve multiple conflict problems. Second, this study examines the effects of the rule-based adaptive controller (RBAC) and the neuro-fuzzy adaptive controller (NFAC) on the adaptation of ITDB. The purpose of this study is to evaluate which of these techniques can yield higher stabilized levels of trustworthiness, data integrity, and data availability in the face of attacks. |
