Data-Driven Network Path Simulation with iBox

While network simulation is widely used for evaluating network protocols and applications, ensuring realism remains a key challenge. There has been much work on simulating network mechanisms faithfully (e.g., links, buffers, etc.), but less attention on the critical task of configuring the simulator to reflect reality.

We present iBox (“Internet in a Box”), which enables data-driven network path simulation, using input/output packet traces gathered at the sender/receiver in the target network to create a model of the end-to-end behaviour of a network path. Our work builds on recent work in this direction [7, 40] and makes three contributions: (1) estimation of a lightweight non-reactive cross-traffic model, (2) estimation of a more powerful reactive cross-traffic model based on Bayesian optimization, and (3) evaluation of iBox in the context of congestion control variants in an Internet research testbed and also controlled experiments with known ground truth.