| 摘要 |
FIELD: physics, computer engineering.SUBSTANCE: invention pertains to computer engineering. Disclosed is a method of providing fault-tolerance of a computer system, based on task replication, self-reconfiguration and self-management of degradation, wherein the computer system includes four numbered computers BMi, where i=1,?,4 is the number of a computer, each connected to of the other computers through its own transmitting interface device with an information broadcast channel, wherein each computer in the operating cycle of the system receives input information, calculates its copy of output information and transmits said copy to every other computer of the system, calculates the correct output information by majoritising its own and all received copies, outputs the majoritisation result to an external medium, compares all available copies of output information with the majoritisation result; each computer further includes first, second, third and fourth program fault counters, first, second, third and fourth program fault seconds counters, first, second, third and fourth program blocking counters, a synchronisation timer, an inter-computer communication (MMO) synchronisation timer, an interruption request generation timer (Tmrgzp), a Tmrgzp preset register, a Tmrgzp counter, a synchronisation timer preset register, a synchronisation timer counter, a MMO preset register, a MMO synchronisation counter and a RFx fixing register.EFFECT: faster operation of the computer system.6 cl, 12 dwg |
