In mountain areas, roads and transport infrastructures are essential from health, economic, social, and environmental and safety points of views but are also highly exposed to natural phenomena. For those critical infrastructures, remote effects due to linking function losses can be very harmful for territories. However, most approaches focus rather on local effects of phenomena and direct vulnerability of people, objects, and infrastructures. At a territorial scale, other techniques allow to consider indirect vulnerabilities and quantify territorial resilience. This paper first describes and then discusses the concept of resilience and then proposes a methodology applied to territorial and transport networks resilience analysis. Network structural properties analysis combines spatial G.I.S analysis, graph theory and produces indicators to (e.g.) evaluate importance, criticality of road sections and also territories accessibility in the context of natural risks. Attractiveness factors on nodes and constraints on edges (physical, exposure to natural phenomena) are valuated using both multi-criteria decision-making methods and rule-based approaches. Then, structural indicators are calculated. The centrality indicator (e.g.) measures the level up to which a road is used to reach any point (somewhere an importance factor). Evolution of the average (mean) topological distance indicator shows how easy or difficult it is to reach a point on the network considering all constraints, including those related to natural phenomena, on networks. GeoGraphLab (GGL) is new improved versatile open-source software which implements this methodology for the analysis of any infrastructure network (transport, energy, and telecommunications).Using those techniques, networks are analyzed under several exposure and failure scenarios in order to assess vulnerability and territorial resilience using relative evolution of indicators.