Mapping The Invisible Internet: Framework and Dataset

TL;DR

This paper introduces a comprehensive dataset and framework for mapping the network topology and traffic of the Invisible Internet Project (I2P), enabling advanced analysis of its decentralized routing and anonymity features.

Contribution

It presents a novel dataset and SWARM-I2P framework for detailed mapping of I2P's network topology, traffic, and node characteristics, which was previously unavailable.

Findings

Mapped over 50,000 I2P nodes with detailed metrics

Provided insights into network topology and geographic distribution

Enabled analysis of routing behaviors and network resilience

Abstract

This article describes a novel dataset that maps the network layer of the Invisible Internet Project (I2P). The data was collected using SWARM-I2P framework, which deployed I2P routers as a network of mapping agents that gather information on the network's topology and traffic over an extended period. The dataset documents over 50,000 nodes, including subsets of high-performance (FastSet) nodes and high-capacity nodes characterized by metrics such as bandwidth, latency, and uptime. It also contains detailed records of network traffic and the geographic distribution of thousands of nodes. Data was collected using a combination of methods, including querying router consoles, analysing the network database (netDb), and passive network monitoring. All node identifiers were anonymized to maintain user privacy. The data is publicly available in CSV and TXT formats on Zenodo, with mapping scripts provided on GitHub. This resource provides a foundational understanding of the decentralized routing behaviours that underpin I2P's anonymity, making it suitable for reuse in analyses of tunnel node selection, anonymity network resilience, and adversarial modelling.