Openflow is one of the representative technologies of SDN, and it is necessary to process a large number of flow table entries when deploying network resources using this technology in cloud computing centers, which is prone to cause problems such as link congestion. The purpose of this paper is to propose a distributed OpenFlow flow table storage architecture (DSFoF). The architecture uses an improved hash algorithm to determine the deployment of flow table items, achieves centralized management of Openflow switches, and achieves large flow table storage and load balancing. The architecture has an average packet loss rate of 2.52% in the range of 6-10Mbps. The bandwidth utilization is basically higher than 80%, and it can effectively avoid node overload within the threshold. The experiments also show that the architecture can handle the network race and improve the throughput, and the task completion is always higher than 95% when the number of pending tasks is within 100, which has a better performance.