The Reference Network
Figure 1 illustrates a network that might be familiar to regular readers.
Figure 1 A service provider network linking two enterprise sites
The center of Figure 1 illustrates a simplified service provider network that interconnects multiple sites of a given enterprise customer. Later, in the code section, I’ll simulate some failure scenarios on the provider network shown in Figure 1. Specifically, I’ll look at two common failures: link failure (the single most common network failure) and congestion. The latter is a very nasty problem because it can cause the network to grind to an ignominious halt. So, the idea is to investigate how the network management system at the bottom of Figure 1 copes with these failures.
As more organizations seek to outsource noncore activities, the role of service providers is becoming increasingly critical. For their part, service providers are also seeking to move up the value chain by offering more sophisticated services such as data backup. Beyond this is an even more specialized provider that offers computational grids. In a sense, this is conceptually a return to the old days of dedicated and centralized IT! The main difference nowadays is that separate organizations now offer the services. The latter are then bought in by the end user organizations.