Research and Application of Multi-processor Fault-Tolerant Algorithms for China Space Station Full Digital Simulation Platform

The improved Byzantine fault tolerance algorithm has been applied to the full digital simulation environment of the China Space Station, achieving a full digital simulation platform that supports multi-processor fault-tolerant for the first time. Based on special needs of China Space Station’s multi-processor architecture, 3 key technologies have been implemented: Firstly, improve the classic Byzantine fault-tolerant algorithm to achieve multi-processor fault-tolerance, even if dual-processor failures occur during the operation of 4 processors, the system can still run safely and stably; Secondly, “Process + Shared Memory” is used to realize the unified management and isolation of data, avoiding the problems of data mutual exclusion and low transmission efficiency in multi-processor simulation; Thirdly, based on the characteristics of data exchange, a new synchronization method has been designed to solve the problem of multi-processor synchronization under fault-tolerant architecture. Ultimately, Implemented a fully digital simulation platform under a multi-processor fault-tolerant architecture of China Space Station, get rid of hardware dependencies. Finally, the application effectiveness of simulation platform is introduced.