Graphs, as noted in the textbook, describe similar situations very well: a member of the internet could draw out every other user and their connections as a graph, and that graph has measurable robustness. For a given service like Amazon AWS or Microsoft Azure, the service's connection to the rest of the web relies on hundreds of smaller connections in a distributed network across the nation.

We want to, using graph theory, study the reliability of those services in terms of bridges or $k$-connectivity, as well as how that could be improved (especially for the consumer market).