Evaluating the effects of cascading failures in a network of critical infrastructures

摘要

Networks of interconnected critical infrastructures are the supporting mechanisms of every industrialised nation. Mutually reliant on each other, their service provisions cross borders. This reliance is also a great weakness. The level of dependence each infrastructure has on another means that a failure has the potential to cascade, resulting in devastating impact on the economy, e-government, defence and society as a whole. Predicting the effects of a cascading failure is a challenge. In this paper, an approach for identifying the effects of a cascading failure is portrayed. A simulation depicting a virtual city is presented, in order to assess the spread of faults originating from a telecommunications infrastructure. Subtle behaviour changes have the potential to spread, with both significant and minor impacts. These variations can be mitigated for using data classification techniques to assess behaviour changes, with an overall accuracy of 85.61% using the TreeC classifier.