End-to-End Delay Testing and Research on the Internet
Zihan Gao, Hua Yu, Xiaozhen Jin, Shiyao Liu, Lei Tang · 2023
According to the principle of network time synchronization, the performance of network time synchronization relies on network transmission delay. Accurate measurement of end-to-end delay in the Internet plays an important role in improving network performance. Data transmission is susceptible to network conditions, and network delay can vary accordingly. In addition, clock deviations and clock frequency differences at both the sending and receiving ends can also affect network delay. To address these issues, this article proposes an end-to-end delay testing method based on time synchronization between the sending and receiving ends. Furthermore, a prediction model based on a backpropagation (BP) neural network is used to predict the end-to-end delay in the Internet. Through training and learning using this method, precise prediction of end-to-end delay can be achieved, laying the foundation for subsequent delay compensation. The feasibility of the algorithm is demonstrated through simulations.