Determine and analyse common cause failures.
A common cause failure is an event that leads to the simultaneous failure of two or more components. For example: two cables in the same duct can both be cut in a single incident; multiple equipment items may be destroyed in a single fire.
For a common cause failure to happen, the affected components must be within range of each other, according to a critical property. For physical failure events such as fire and flooding, this property is geographical proximity: the components must be sufficiently close to be affected simultaneously. For configuration mistakes it is the similarity in maintenance procedures. For software bugs it is whether related firmware versions are used, regardless of geographical distance. Other events may have different critical properties.
For each failure scenario, the critical property has a maximum effect distance. Two equipment items can only be affected by a minor fire when they are in the same room; for a major fire the effect distance is larger, but still limited to perhaps a single building. Flooding has a much larger effect area, and two components must be further apart to be immune from flooding as their common failure cause.
In stage 3 you will make groups of components that fall within the same range of a critical property. You will do this for each vulnerability separately. For each cluster you will then assess the Frequency and Impact of a common cause failure affecting the components in that cluster. The clusters and their assessments will be recorded in the Raster tool. The result is an improved and refined risk assessment.
The Common Cause Failures Analysis stage consists of the following steps: