A Fine-Grained Network Congestion Detection Based on Flow Watermarking

With the rapid development of network, how to effectively reduce dynamic delay and improve performance network is an important challenging problem. Specifically, congestion one key factors that hurt performance, so real-time detection critical for recovering failure quickly. Current research in mainly faces problems occupying extra bandwidth, decreasing ratio effective payload packet, increasing burden switches, etc. In this paper, we apply flow watermarking propose a fine-grained method based on watermarking. We also combine it with eBPF (extended Berkeley Packet Filter) detection. Theoretical analysis experimental results show changes status can be reflected accurately watermark decoding information. The detect small time scale realize low-overhead easily deployed