RTT Algorithm Optimization with Out-of-Order Retransmission Detection
Wei Jiang, Bin Zhang, Wang Wenyong, Zhu Qixun · 2024
Traditional Round-Trip Time (RTT) algorithms often struggle with maintaining accuracy under conditions of packet retransmissions and disorder, resulting in RTT estimations that fall outside the expected range. This can lead to misinterpretations of network health and performance. In response to this challenge, our research introduces an enhanced RTT calculation method designed for implementation within network devices. Our approach involves a bidirectional monitoring mechanism for detecting packet retransmissions during the process of traffic interception. By doing so, we aim to refine the precision of RTT estimations through meticulous packet inspection. Our novel algorithm is rigorously tested under simulated environments that mimic fluctuating network conditions. The outcomes clearly indicate that our methodology successfully screens out anomalous RTT values, which are typically introduced by packet retransmissions or disorder. Consequently, this leads to a substantial enhancement in the accuracy of RTT calculations. Our findings suggest that the adoption of this optimized RTT algorithm could significantly improve the reliability of network performance monitoring systems, making it a promising solution for real-world network management applications. The comparative experimental results of RTT optimization before and after simulating abnormal network environments show that the peak value of sudden abnormal RTT over the family side and the network side after optimization is reduced by about 75.8% and 99.5%, respectively.