Patterns for Fault Tolerant Software

Chapter 7. Error Mitigation Patterns

The patterns in the previous chapter discuss error recovery methods in which the state of the executing system is changed to permit the system to continue. The patterns in this chapter describe ways to mitigate the effects of the error without changing the application or system state. The mitigation techniques mask errors and compensate for their effects.

Errors are manifest as either an incorrect value or system state; or as a missed timing requirement or constraint. These patterns describe ways to mitigate the effects of both of these types of errors. When the error is of a value or state kind, the mechanisms described in these patterns correct or change the erroneous value and state, or the underlying fault, so that processing can continue from the point at which the error was detected. This differs from the patterns presented in the previous chapter that are concerned with taking the system to a new state from which execution can continue without the error.

When the error is missed timing or other issues as the result of an excess of workload, the patterns in this chapter present proven techniques to mitigate the effects by giving the system the means to gracefully survive the excess. The underlying faults may not be directly addressed because they might not be local to the overloaded system. For example, the performance error might be the result of an excess of work arriving at the system from outside. The patterns instead design mechanisms ...