Design Tactics - Availability

Issue a ping and expects to receive back an echo, within a predefined time.

Listen to a heartbeat message periodically emitted periodically by a component.

Raise exception when one of the fault classes is recognized.

Run processes running on redundant processors, where each process takes equivalent input and compute a simple output value that is sent to a voter.

Have redundant components respond to events in parallel, they are all in the same state.

Have one component (the primary) respond to events and inform the other components (the standbys) of state updates they must make.

Have a spare computing platform configured to replace many different failed components.

Run a previously failed component in 'shadow mode' for a short time to make sure that it works.

Restore the component to have its state upgraded before its return to service.

Record a consistent state, roleback to it on failure.

Remove a component from operation to prevent anticipated failures.

Bundle several sequential steps such that the entire bundle can be undone at once.

Have a monitoring process delete the nonperforming process and create a new instance of it.