The Study of a Distributed Networking Method Based on an Improved Raft Protocol

To address issues such as central server dependency, harsh environments, and network instability during the networked upgrade of industrial control devices, this paper proposes a distributed self-organizing network system based on the Raft consensus algorithm. By comparing and analyzing the feasibility, advantages, and disadvantages of the Paxos and Raft protocols, the Raft protocol was selected for constructing the distributed system. To overcome the limitations of the Raft algorithm in the leader election process, a leader election algorithm based on node priority was proposed and implemented to enhance the performance of the distributed system under high load and unstable network conditions. Simulation experiments were conducted, and the results demonstrate that the optimized algorithm proposed in this paper reduces the average write latency by 18.3%, increases the number of logs processed per unit of time by 27.2%, and improves the system throughput per unit of time by 39.9%, compared to the original Raft algorithm.