Two types of networks

  • Models

Networks are strange things, Mr. Node thought. It seems like there are two types of them. One for forest fires and one for logistics. The networks that arise when you strive for an average R0 value create the dynamics of the forest fire: By R0 = 1, most trees are too far apart to light each other, at R0 = 2 groups of trees are created that can light each other, at R0 = 3 those groups become larger and at R0 = 4 one burning tree can burn the whole forest. All trees are connected when they are within range of each other for transmitting fire. But trees are standing still. Forest fires move between immobilia. Logistics often transports moving targets from sending moving targets to receiving moving targets. And wherever people move, those who are infective may be among them.

The latter results in a different type of infectious networks. Much more dynamic and therefore much more dangerous when occurring in epidemics. It ensures that a new pandemic can spread all over the world within a handful of weeks. Their architecture can be simulated by building a slight preference (in who is looking for connections) for linking to who already have have links (preferential attachment). I show examples of the two types of network architecture in Fig. 1.

The logistics architecture is characterized by hubs that encourage uniformity and are therefore efficient but also vulnerable. Forest fire architectures are less efficient, but also less vulnerable.

We know how we can promote the formation of those architectures, how we can mix them, and how we can hurt them, although this is more difficult with the forest fire architectures. It is interesting that the initially deliberate forest fire architecture of the internet has been transforming into a logistics architecture as a result of its commercial success.

Mister Node does notice that taking measures against the spread of the COVID-19 virus (initially a logistic affair) often results in a move towards a forest fire architecture.