Simulating Network Paths with Recurrent Buffering Units

We developed an end-to-end network path simulation model that embeds the semantics of a physical network, leveraging domain-specific insights from physical network paths and explicitly modelling unobservable cross-traffic using a new RNN-style architecture called Recurrent Buffering Unit (RBU). Our method beats the state-of-the-art (iBoxNet) on real-network traces and surpasses state-of-the-art generative models such as GANs and Transformers, which suffer from several drawbacks in this setting. Furthermore, our method accounts for complex behaviour such as packet reordering which is not possible with iBoxNet.

[Arxiv]