Research on Adaptive Congestion Avoidance Algorithm Based on RTT Congestion Signal

Traditional TCP congestion control algorithms, such as TCP Reno and TCP New Reno, mainly rely on packet loss signals to adjust the congestion window, which will lead to low utilization of network resources to a certain extent, especially in the network environment with large RTT. To solve this problem, this paper proposes a congestion avoidance algorithm based on RTT adaptive congestion window. By introducing RTT as a key factor of congestion window adjustment, TCP congestion control is more refined and network resource utilization is improved. The algorithm determines the current network environment by RTT and changes the congestion window according to different network environments to realize the adaptive change of the congestion window. In this way, the algorithm can respond to the changes of the network more timely and accurately, thus improving the performance of the whole network. The simulation results show that the improved algorithm reduces the data transmission delay and improves the congestion window and throughput significantly.