A Method for Initial Synchronization of Multi-computer Hot Backup Computers

Abstract

The invention discloses an initial synchronization method for multiple hot backup computers. The initial synchronization method comprises the steps that each hot backup computer receives the same externally-inputted synchronous clock signal; each hot backup computer finishes the initialization work and then transmits information A to other units, wherein the contents of the information A are that a local computer has synchronous working conditions; a certain hot backup computer receives the information A transmitted by all other hot backup computers and transmits information B to other hot backup computers, wherein the contents of the information B are that when a rising edge of next synchronous clock signal comes in, the current hot backup computer starts to work; after a certain computer receives the information B, the certain computer transmits the information B to other computers; if a certain hot backup computer does not receive the information A transmitted by other computers within continuous N (N is a preset value) synchronous clock cycles, the certain hot backup computer transmits the information B to the other computers. According to the initial synchronization method disclosed by the invention, initial synchronization of the multiple hot backup computers can be realized, and the reliability and the independent operation ability of the hot backup computer are improved.

Description

technical field [0001] The invention relates to an initial synchronization method for a multi-computer hot backup computer, which is suitable for an on-board computer with higher initial synchronization requirements or autonomous operation capability for multiple computers. Background technique [0002] At present, the synchronization of the three machines in our country is realized by the unified control period signal provided by the fault-tolerant board, and the control period signal is completely asynchronous with the clock of the three machines. This design can realize the synchronization of the three machines, but it cannot guarantee the initial synchronization of the three machines (that is, the three machines enter the control cycle for the first time to interrupt and start task scheduling and execution at the same time). The specific reasons are as follows: [0003] There is a difference in the power-on / reset initialization time of the three machines, resulting in th...

AI Technical Summary

Problems solved by technology

[0003]There is a difference in the power-on / reset initialization time of the three machines, resulting in the risk that the three machines cannot be initially synchronized after power-on / reset initialization is completed

[0004] The reset signal interface circuits of the three machines on the fault-tolerant board are completely independent, and there are differences in the recognition of the same reset signal by different interface circuits, resulting in differences in the end times of the reset signals of the three machines (when the three machines are powered on at the same time, the difference in the end time of the power-on reset is about 10ms, The time difference between remote control and watchdog reset end time is about 1us)

[0005] When the three machines are powered on at the same time, the power modules of the three machines are different, resulting in a large difference in the power establishment time of the three machines, and there is a risk that the initial synchronization cannot be powered on

When the three machines are powered on sequentially, the three machines cannot be powered on for initial synchronization. In this way, it is often necessary to reissue the remote control reset command of the three machines to achieve the initial synchronization of the three machines. However, the remote reset also has the risk of being unable to perform initial synchronization, and some On-orbit missions are not allowed to send remote resets or do not have the conditions to send remote resets

[0006] As for the initial out-of-sync problem of the hot backup computer, the hot backup computer of the previous model usually handles this problem through the application layer, and adopts various methods to tolerate the initial out-of-sync of the three machines, or Avoid this problem by reissuing remote control reset (there is a certain probability that this problem cannot be avoided)

However, there are still defects in the synchronization ability and reliability of the three-machine hot backup computer, and at the same time, the independent operation ability of the product is also reduced, and it is difficult to apply to the design reliability requirements and autonomous operation ability of several major models of three-machine or multi-machine hot backup computers in the future.

Images