An elementary Ising spin model is proposed for demonstrating cascading failures (breakdowns, blackouts, collapses, avalanches, ...) that can occur in realistic networks for distribution and delivery by suppliers to consumers. A ferromagnetic Hamiltonian with quenched random fields results from policies that maximize the gap between demand and delivery. Such policies can arise in a competitive market where firms artificially create new demand, or in a solidary environment where too high a demand cannot reasonably be met. Network failure in the context of a policy of solidarity is possible when an initially active state becomes metastable and decays to a stable inactive state. We explore the characteristics of the demand and delivery, as well as the topological properties, which make the distribution network susceptible of failure. An effective temperature is defined, which governs the strength of the activity fluctuations which can induce a collapse. Numerical results, obtained by Monte Carlo simulations of the model on (mainly) scale-free networks, are supplemented with analytic mean-field approximations to the geometrical random field fluctuations and the thermal spin fluctuations. The role of hubs versus poorly connected nodes in initiating the breakdown of network activity is illustrated and related to model parameters.Version: April 20, 2018 keywords: Ising model, random fields, complex scale-free distribution network, failure, blackout, breakdown, collapse, avalanche, cascade MotivationStatistical physics studies of cooperative phenomena on random graphs have attracted a lot of new attention and undergone impressive new development since it has become clear that many reallife interconnected structures are well approximated by random scale-free networks [1,2,3,4,5]. One can say that a paradigm shift is occurring from studies of models for critical phenomena on (Bravais) lattices to studies in which such models are defined on random networks. To some extent this paradigm shift resembles that from Euclidean geometry to fractal geometry, in the modeling of various natural phenomena, but scale-free networks are specific in that they are characterized by the presence of a small but important set of hubs. The hubs are highly connected nodes which typically have a large influence on the operation and coherence of the structure.When we are concerned with the distribution of electricity, or the production and sale of material goods by commercial centers to consumers, or the delivery of the daily mail ..., we can * Dedicated to Professor Dr. David Sherrington on the occasion of his 70th birthday. 1 envisage distribution centers as nodes and consumers as links on a network. The consumers have certain demands and the distribution centers certain deliverables. The distribution centers can be active or inactive depending on internal conditions and on external criteria such as the demands of consumers that are linked to them and the status of other nearby centers. The occurrence of hubs in such networks is rathe...