Abstract: In industrial control, the Paxos algorithm often ensures data consistency but lacks strict determinism in consensus completion time and has high communication overhead, making it difficult to meet the real-time and deterministic requirements of Time-Sensitive Network (TSN). This paper proposed TSPaxos, an efficient Paxos algorithm for TSN. By utilizing TSN's high-precision time synchronization capability, TSN dynamically allocates transmission time slots for different consensus stages, enhancing the determinism of consensus completion time. Leveraging TSN's deterministic transmission characteristics, an implicit acknowledgment mechanism was introduced, triggering explicit feedback only when a node rejects a message or needs to send historical proposals. Theoretical analysis showed this mechanism can reduce consensus messages by over one-third. Based on this, a dynamic time slot adjustment mechanism was designed to improve the robustness of the algorithm. Experimental results show that compared with three typical optimized algorithms, the TSPaxos algorithm achieves performance improvements of over 32.53%, 58.78%, and 97.39% in terms of consensus message count, average consensus latency, and standard deviation of consensus completion time, respectively.

Key words: consensus algorithm, Paxos, Time-Sensitive Network (TSN), industrial control system, implicit acknowledgement

摘要： 工控领域通常依赖Paxos算法实现数据一致性。然而，Paxos在共识完成时间上缺乏严格确定性，且通信开销较高，难以满足时间敏感网络(Time-Sensitive Network, TSN)的实时性和确定性要求。为此，提出一种适用于TSN的Paxos高效算法TSPaxos。算法利用TSN高精度时间同步能力，动态分配不同共识阶段的传输时隙，增强共识完成时间确定性；利用TSN确定性传输特性，引入隐式确认机制，仅当节点拒绝消息或需发送历史提案时，才触发显式反馈，理论分析表明，该机制能够减少超过三分之一的共识消息数。在此基础上，设计动态时隙调整机制，提升算法鲁棒性。实验结果表明，相比三种典型优化算法，TSPaxos算法在共识消息数、平均共识延迟以及共识完成时间标准差方面性能提升达到32.53%、58.78%以及97.39%以上。

关键词: 共识算法, Paxos, 时间敏感网络, 工业控制系统, 隐式确认