Energy-Efficient Deep Learning for Traffic Classification on Microcontrollers

TL;DR

This paper introduces a lightweight, energy-efficient deep learning model for traffic classification on microcontrollers, achieving high accuracy and low power consumption suitable for IoT security applications.

Contribution

It develops a hardware-aware neural architecture search optimized 1D-CNN that balances accuracy and efficiency for deployment on resource-constrained devices.

Findings

Achieves 96.59% accuracy with 88.26K parameters.

Quantized model drops only 1-2% accuracy.

Inference latency under 32ms on high-performance microcontrollers.

Abstract

In this paper, we present a practical deep learning (DL) approach for energy-efficient traffic classification (TC) on resource-limited microcontrollers, which are widely used in IoT-based smart systems and communication networks. Our objective is to balance accuracy, computational efficiency, and real-world deployability. To that end, we develop a lightweight 1D-CNN, optimized via hardware-aware neural architecture search (HW-NAS), which achieves 96.59% accuracy on the ISCX VPN-NonVPN dataset with only 88.26K parameters, a 20.12K maximum tensor size, and 10.08M floating-point operations (FLOPs). Moreover, it generalizes across various TC tasks, with accuracies ranging from 94% to 99%. To enable deployment, the model is quantized to INT8, suffering only a marginal 1-2% accuracy drop relative to its Float32 counterpart. We evaluate real-world inference performance on two microcontrollers: the high-performance STM32F746G-DISCO and the cost-sensitive Nucleo-F401RE. The deployed model achieves inference latencies of 31.43ms and 115.40ms, with energy consumption of 7.86 mJ and 29.10 mJ per inference, respectively. These results demonstrate the feasibility of on-device encrypted traffic analysis, paving the way for scalable, low-power IoT security solutions.