Using realistic factors to simulate catastrophic congestion events in a network
Kevin L. Mills, Christopher E. Dabrowski
With the rapid growth of the Internet, there has been increased interest in the use of computer models to study the dynamics of communication networks in the research literature. An important example of this has been the study of dramatic, but relatively infrequent, events that result in abrupt, and often catastrophic, failures in the network due to congestion. With few exceptions, the computer models used in these previous studies have been abstract graphs that include simplified models of such important network factors as router speeds and congestion control procedures. Here, we modify this typical approach, adding realistic network factors which can be varied, including router classes, variable router speeds, flows, TCP, sources and receivers, and packet dropping to a graph model of a single Internet Service Provider (ISP) network that can have more than a quarter million members. For each valid configuration of realistic factors, congestion is then gradually increased. While there are realistic network models reported in the literature, to our knowledge none of these have been used to study catastrophic failures. Further, none of these network models have realistic factors which can be varied. We show that the addition of realistic network factors to our model of an ISP network can mitigate catastrophic events. With the addition of variable router speeds or TCP, a transition to a congested state, where all routers are congested in a single ISP network does not take place. Yet, as congestion spreads, ultimately the operation of the ISP network appears to decline, along with the ability of its nodes to communicate.