Real-Time Fault-Tolerant Scheduling for Distributed Systems Based on Decentralized-Backup

In scenarios where time constraints are required to be precise, the system needs to meet the real-time task constraints even when a processor fails. Addressing the issue that existing fault-tolerant fair scheduling algorithms cannot completely prevent task rejections, this paper designs the DB-FTFS (Decentralized-backup Fault Tolerant Fair Scheduling) scheduling algorithm based on a decentralized-backup. The algorithm prevents the system from entering an unsafe state by determining the minimum requirements of tasks and ensuring the execution of these requirements. The Decentralized-backup designed in this paper is used to provide redundant backups for tasks, ensuring that tasks have access to adequate resources even when potential errors are present in the system. Experimental results show that after a permanent fault occurs in the kernel, the DB-FTFS algorithm can prevent task rejections caused by the fault, reducing the number of rejected tasks to 0.