Fault tolerance: Difference between revisions
Jump to navigation
Jump to search
imported>Howard C. Berkowitz No edit summary |
imported>Howard C. Berkowitz No edit summary |
||
| Line 2: | Line 2: | ||
In engineering, '''fault tolerance''' is a characteristic of a system that can have one or more subcomponents fail without the entire system failing. This does not mean that the system has no single point of failure, but that at least some parts can fail and have continued operation; a system might be robust to a component failure, but have no backup for complete physical destruction. | In engineering, '''fault tolerance''' is a characteristic of a system that can have one or more subcomponents fail without the entire system failing. This does not mean that the system has no single point of failure, but that at least some parts can fail and have continued operation; a system might be robust to a component failure, but have no backup for complete physical destruction. | ||
With the loss of some components, there might be a loss of performance or functionality, or the system could be engineered with reserve capacity such that some failures would not be noticed by users. | With the loss of some components, there might be a loss of performance or functionality, or the system could be engineered with reserve capacity such that some failures would not be noticed by users. | ||
Simple duplication of components does not make a system fault-tolerant; there needs to be an intelligent mechanism for spreading the work around the failure. | Simple duplication of components does not make a system fault-tolerant; there needs to be an intelligent mechanism for spreading the work around the failure. | ||
Revision as of 17:29, 5 June 2009
In engineering, fault tolerance is a characteristic of a system that can have one or more subcomponents fail without the entire system failing. This does not mean that the system has no single point of failure, but that at least some parts can fail and have continued operation; a system might be robust to a component failure, but have no backup for complete physical destruction.
With the loss of some components, there might be a loss of performance or functionality, or the system could be engineered with reserve capacity such that some failures would not be noticed by users.
Simple duplication of components does not make a system fault-tolerant; there needs to be an intelligent mechanism for spreading the work around the failure.