Modeling Network Resilience and Utility of Services

We present methods for calculating network resilience measures based on classical reliability polynomials. Network topology and redundancy of critical elements are essential aspects in the modeling. Three typical case studies are discussed and illustrated with a small real-world infrastructure network. The first two examples demonstrate average resilience of nodes in the network and effects of multiplying application services. The third case study demonstrates how networked service level and physical network level can be combined to achieve an overall utility measure. Service layers are aggregated to a matrix describing average values of services. This matrix is further aggregated with the corresponding reliability matrix of the same physical network by taking the of Hadamard product. Similar methods can be used for modeling both cyber-attacks and conventional technical failures.