March 2003
Intermediate to advanced
912 pages
27h 17m
English
If a program runs on a single machine and a failure occurs we assume that all the memory is lost; that is, all the data structures of the program are lost. If no externally visible effects have been caused by the program it is as though it had never run and it can simply be restarted. In practice, every program will eventually perform output or cause some change of permanent state or communicate an intermediate result. If a crash occurs during a program run and after such an externally visible action, we consider how to restart after the crash.
An aspect of system design is the extent to which an application is offered support to recover from system crashes. Another name for crash resilience is ...
Read now
Unlock full access