An End-to-end Flow Control Method Based on DQN
Guichao Gao, Rong Jin · 2022 International Conference on Big Data, Information and Computer Network (BDICN) · 2022
We propose an end-to-end flow control method based on DQN (Deep-Q-Network). The flow control method is intelligent and flow-based, it is suitable for those DCNs (Data Center Network) based on SDN (Software Define Network) architecture. The system architecture is analyzed and an improved DQN algorithm is proposed. The training results show that the method has good convergence speed and convergence effect. The final test shows that the DQN-based end-to-end flow control method can effectively achieve DCN flow control by allocating the flow rates globally, and can improve the link utilization rate of the entire network as much as possible.