HH-NIDS: Heterogeneous Hardware-Based Network Intrusion Detection Framework for IoT Security

This study proposes a heterogeneous hardware-based framework for network intrusion detection using lightweight artificial neural models. With the increase in volume of exchanged data, IoT networks’ security has become crucial issue. Anomaly-based systems (IDS) machine learning have recently gained increased popularity due to their generation’s ability detect unseen attacks. However, deployment anomaly-based AI-assisted IDS devices is computationally expensive. A high-performance and ultra-low power consumption proposed evaluated this paper. The achieved highest accuracy 98.57% 99.66% on UNSW-NB15 IoT-23 datasets, respectively. inference engine MAX78000EVKIT AI-microcontroller 11.3 times faster than Intel Core i7-9750H 2.6 GHz 21.3 NVIDIA GeForce GTX 1650 graphics cards, when drawn was 18mW. In addition, pipelined design PYNQ-Z2 SoC FPGA board with Xilinx Zynq xc7z020-1clg400c device optimised run at on-chip frequency (100 MHz), which shows speedup 53.5 compared MAX78000EVKIT.