As described in the first section of this chapter, a distributed system is expected to be more reliable than a centralized system. It seems that the existence of multiple instances increases the level of system reliability. On the flip side, few design challenges on loosely coupled systems have the probability of unreliable characteristics in distributed applications.

A software reliability curve is pictorially represented as follows:

With the inference of the preceding graph, software reliability is defined as the probability of the failure rate in the y axis against the specific time period in the x axis for a particular distributed environment.

By rolling higher reliability in an enterprise application, it is assured that you will be able to avoid/tolerate system faults, tolerate faults, and detect/recover from the identified faults.