Redundancy management and switching method of dual-computer redundancy system

Abstract

The invention discloses a redundancy management and switching method of a dual-computer redundancy system, which comprises two computers. Software modules in each computer are divided into a redundancy management module for coordinating and managing redundancy of a complete computer, a plurality of interface modules for supplying external services and a plurality of kernel modules for achieving specific operations. System redundant states are arranged in the redundancy management module, and each interface module and each kernel module are respectively provided with a local redundant state copy corresponding to the system redundant states. The mode of software module classification and arrangement of the redundant state copies is used, thereby being capable of reducing lock contention and false sharing called by multiple threads, improving performance of the dual-computer redundancy system, and particularly being adapted to multi-core and multi-central-processing-unit (CPU) computer environment.

Description

Technical field [0001] The invention belongs to the field of industrial automation. Specifically, it relates to a dual-computer redundancy system, and more specifically, to a redundancy management and switching method of a dual-computer redundancy system. Background technique [0002] Redundancy technology is a commonly used technology in computer system reliability design, and it is an effective method to improve computer system reliability. In the industrial automation system, in order to make the system run stably and reliably for a long time, some form of redundant system is often used, among which the dual redundant system is the most used. In a dual-computer redundant system, there are two computers with the same hardware configuration, which are connected through a network, run the same software, and realize data sharing and synchronization. When running, one of the two computers serves as a working machine, and the other as a backup machine. If the working machine fails...

AI Technical Summary

Problems solved by technology

[0004] Since there is only one redundant state in the computer, when external programs and internal processes read the state, they will add exclusive locks or shared locks to the state as needed, resulting in lock competition

Since there are often multiple processes reading the redundant state at the same time, the lock competition when multi-threaded calls is very serious, resulting in a serious decline in system performance

On the other hand, when the computer is in a multi-core and multi-CPU environment, when there is only one redundant state for reading, it is very easy to cause false memory sharing, which further reduces system performance.

Images