NetSSM: Multi-Flow and State-Aware Network Trace Generation using State-Space Models

2025-03-28

AI Summary

Existing network trace generation methods struggle to model stateful interactions across multi-flow sessions and lack robust, application-oriented evaluation. This paper introduces the first packet-level network trace generation framework based on State Space Models (SSMs), enabling joint learning of multi-flow coordination and session-level state evolution—previously unachieved. Our approach integrates protocol-aware representation learning, multi-flow temporal modeling, and long-range dependency capture, supporting generation of ultra-long traces (up to 8–78× longer than mainstream Transformer-based methods). Extensive evaluations across multiple benchmarks demonstrate that our generated traffic significantly outperforms existing methods in protocol semantic compliance, flow- and session-level statistical fidelity, and application-layer behavioral accuracy. The framework thus establishes a reliable, low-resource data foundation for network research and system validation.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Web Mining and Content Analysis: Web data generation and simulation
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications

Abstract

Access to raw network traffic data is essential for many computer networking tasks, from traffic modeling to performance evaluation. Unfortunately, this data is scarce due to high collection costs and governance rules. Previous efforts explore this challenge by generating synthetic network data, but fail to reliably handle multi-flow sessions, struggle to reason about stateful communication in moderate to long-duration network sessions, and lack robust evaluations tied to real-world utility. We propose a new method based on state-space models called NetSSM that generates raw network traffic at the packet-level granularity. Our approach captures interactions between multiple, interleaved flows -- an objective unexplored in prior work -- and effectively reasons about flow-state in sessions to capture traffic characteristics. NetSSM accomplishes this by learning from and producing traces 8x and 78x longer than existing transformer-based approaches. Evaluation results show that our method generates high-fidelity traces that outperform prior efforts in existing benchmarks. We also find that NetSSM's traces have high semantic similarity to real network data regarding compliance with standard protocol requirements and flow and session-level traffic characteristics.

Problem

• Generates synthetic multi-flow network traffic data
• Models stateful communication in long-duration sessions
• Improves fidelity and utility of synthetic traces

Innovation

• State-space models for packet-level traffic generation
• Handles multi-flow and stateful communication effectively
• Generates long, high-fidelity traces with semantic similarity