NetReplica: A Programmable Substrate for Bottleneck-Centric Network Data Generation

Abstract

The behavior of Internet applications is shaped by congestion dynamics at bottleneck links, yet data capturing application behavior across diverse bottleneck regimes remains scarce. Bridging this gap requires a data-generation substrate that simultaneously provides controllability, composability, fidelity, and replicability, capabilities that existing approaches struggle to achieve together. This paper introduces NetReplica, a programmable substrate for bottleneck-centric data generation guided by progressive disaggregation. NetReplica (i) decouples bottleneck intent from execution, (ii) separates static bottleneck attributes from dynamic congestion pressure, and (iii) disaggregates observed demand dynamics from their original trace context through Cross-Traffic Profiles (CTPs). CTPs transform passive packet traces into reusable, composable pressure signals that can be selected and transformed to specify dynamic bottleneck behavior. Our evaluation shows that NetReplica provides all four capabilities simultaneously and, in an ABR case study, generates datasets that remain realistic while expanding coverage of underrepresented regimes, improving the performance of trained models. In particular, NetReplica reduces transmission-time prediction error for the well-explored Fugu model by up to 47%. Together, these results demonstrate that NetReplica is a practical, fully programmable bottleneck-centric data-generation substrate for developing production-ready protocols, applications, and learning artifacts.