# CNN-based IoT Device Identification: A Comparative Study on Payload vs. Fingerprint

arXiv: https://arxiv.org/abs/2304.13894 · HTML: https://mlsanity.com/paper/2304.13894/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 3 of its 13 references that are in the map.
- First posted on arXiv 27 Apr 2023.

## Builds on (most cited first, 3 of 3)

- [Auto-Keras: An Efficient Neural Architecture Search System](https://mlsanity.com/paper/1806.10282/) (2018)
- [Deep Packet: A Novel Approach For Encrypted Traffic Classification Using Deep Learning](https://mlsanity.com/paper/1709.02656/) (2017)
- [IoT Sentinel: Automated Device-Type Identification for Security Enforcement in IoT](https://mlsanity.com/paper/1611.04880/) (2016)

## Abstract (arXiv)

The proliferation of the Internet of Things (IoT) has introduced a massive influx of devices into the market, bringing with them significant security vulnerabilities. In this diverse ecosystem, robust IoT device identification is a critical preventive measure for network security and vulnerability management. This study proposes a deep learning-based method to identify IoT devices using the Aalto dataset. We employ Convolutional Neural Networks (CNN) to classify devices by converting network packet payloads into pseudo-images. Furthermore, we compare the performance of this payload-based approach against a feature-based fingerprinting method. Our results indicate that while the fingerprint-based method is significantly faster (approximately 10x), the payload-based image classification achieves comparable accuracy, highlighting the trade-offs between computational efficiency and data granularity in IoT security.