Knowledge-based key-hole monitoring of users for security control in transaction processing systems

This paper considers the interrelationships between knowledge-based security, concurrency and recoverability in on-line database transaction processing systems (OLTP). It describes how OLTP can permit the users to carry out the desired actions, and concurrently activate scrutinizing transactions that identify the users with respect to their biometrical or other characteristics stored in a secure knowledge database. A finite state machine model is used to design protocols for recovery, under security breaches. It is shown that the best way to achieve concurrency and recovery is through shadowing and strict scheduling in conjunction with certification locking or timestamping. One must, however, bear in mind that security and concurrency have conflicting requirements and at best one can only achieve a compromise between them.