RISK-MEANINGFUL DATABASES

Abstract

Risk-meaningful databases are central to one of three important approaches to risk elimination in systems, namely use of both a risk monitoring procedure for detection of risk, and a response procedure for either eliminating or reducing the risk detected. The risk monitoring procedure uses violations of risk-meaningful database constraints to signal the presence of risk. The degree to which risk is can be detected depends on a cumulative measure of the risk-meaningful database constraints available. This cumulative measure may be used in an extended version of the risk equation for systems, in order to quantify the relationship between throughput capacity and risk that has been partly reduced or eliminated by means of a risk monitoring procedure. Two types of risk-meaningful databases are identified, namely active rm-databases and passive rm-databases. Passive rm-databases are mostly constant in size, they store data about hazardous entities, and can have any of the structures commonly found in databases. In contrast, active rm-databases grow continually, they store incoming data from monitoring units and sensors, they tend to have a more restricted structure, and the rm-database constraints used for detecting risk depend heavily on co-relationships, that is, relationships arising from entity or event relations that have non-primary key attributes in common.