Today’s IT systems provide business services by means of applications and middleware that depend on highly complex networks of interrelated hardware and software components. IT component failures can impact multiple mission-critical business services, and can lead to cascading chains of failures in other components. Researchers at IBM Research - Zurich are developing models and tools that analyze information from service records, system logfiles, and configuration data to build statistical models of the risks to business service availability. These models are based on the reliabilities of underlying IT components and their interdependencies. The tools interface with technologies for automated discovery of service dependencies to maintain an up-to-the-minute picture of an actual IT configuration. The statistical predictions from the tools help managers to better respond to risk events as they occur, based on their likely impact on business services, as well as to assess the risk impacts of reconfigurations of the IT environment.
A structured approach to IT risk modeling can be applied to either a top-down or a bottom-up view of risk within the organization.