Читать книгу From Traditional Fault Tolerance to Blockchain - Wenbing Zhao - Страница 40

Checkpoint-based protocols only ensure to recover the system up to the most recent consistent global state that has been recorded and all executions happened afterwards, if any, are lost. Logging can be used to recover the system to the state right before the failure, provided that all events (that could potentially change the state of the processes) are logged and the log is available upon recovery. This is what is referred to as the piecewise deterministic assumption [21]. According to this assumption, all nondeterministic events can be identified and sufficient information (referred to as a determinant [1]) must be logged for each event. The most obvious example of nondeterministic events is the receiving of a message. Other examples include system calls, timeouts, and the receipt of interrupts. In this chapter, we typically assume that the only nondeterministic events are the receiving of a message. Note that the sending of a message is not a deterministic event, i.e., it is determined by a nondeterministic event or the initial state of the process [7].