A Crowdsensing Intrusion Detection Dataset For Decentralized Federated Learning Models

TL;DR

This paper presents a large, diverse dataset for IoT malware detection and evaluates decentralized federated learning models, demonstrating their effectiveness and data privacy advantages over centralized approaches.

Contribution

It introduces a comprehensive IoT crowdsensing malware dataset and provides an experimental comparison of decentralized federated learning with traditional methods.

Findings

DFL maintains competitive performance with data locality.

DFL outperforms CFL in most experimental settings.

The dataset enables advanced security research for IoT environments.

Abstract

This paper introduces a dataset and an experimental study on Decentralized Federated Learning (DFL) for Internet of Things (IoT) crowdsensing malware detection. The dataset comprises behavioral records from benign and eight malware attacks. A total of 21,582,484 original records were collected from system calls, file system activities, resource usage, kernel events, input/output events, and network records. These records were aggregated into 30-second windows, resulting in 342,106 data records used for model training and evaluation. Experiments on the DFL platform compare traditional Machine Learning (ML), Centralized Federated Learning (CFL), and DFL across different node counts, topologies, and data distributions. Results show that DFL maintains competitive performance while preserving data locality, outperforming CFL in most settings. This dataset provides a solid foundation for studying the security of IoT crowdsensing environments.