Fault Tolerance
Fault tolerance is built into VOS from the bottom up. On a hardware level, major devices are run in lock stepped duplex mode, meaning that there are two identical devices performing the same action at the same time. (In addition, each device, or board, is also duplexed in order to identify internal board failures at a hardware level, which is why Stratus hardware can be defined as "lock stepped".) These boards are actively monitored by the operating system which can correct any minor inconsistencies (such as bad disk-writes or reads). Any boards which report an unacceptable number of faults are removed from service by the system; the duplexed board will continue operation until the problem is resolved via a hot-fix. This includes CPUs, disk drives, and any other device that can logically be duplexed (which by definition, excludes communications devices). The system will continue processing as normal and will automatically raise a fault ticket with Stratus Customer Service via RSN (the Remote Service Network). Stratus Customer Service will then dial into the system using RSN to investigate the problem and dispatch replacement parts.
Typically a modern Stratus ftServer or Continuum can expect to see 99.999% uptime. This is not to say that the applications running on these systems will achieve this level of uptime, only that the operating system will not crash due to a simplexed hardware failure.
Read more about this topic: Stratus VOS
Famous quotes containing the words fault and/or tolerance:
“The Law is the true embodiment
Of everything that’s excellent.
It has no kind of fault or flaw,
And I, my Lords, embody the Law.”
—Sir William Schwenck Gilbert (1836–1911)
“Advocating the mere tolerance of difference between women is the grossest reformism. It is a total denial of the creative function of difference in our lives. Difference must be not merely tolerated, but seen as a fund of necessary polarities between which our creativity can spark like a dialectic.”
—Audre Lorde (1934–1992)