Hybrid Deep Learning Model for Detecting DDoS Attacks in IoT Networks
V. Jyothsna, Chaithanya Kumar Reddy Vardhireddy, Hemasree Thangella, Kartheek Sarangula, Roshini Tamidilapati, Bhasha Pydala
Abstract
As the number of internet connected devices has surpassed tens of billions, the era of the "Internet-of-Things" (IoT) is here.These days, a vast array of products seamlessly integrate the internet, from small devices like smartwatches to more intricate systems like smart grids, smart transit networks, and smart cities. Apart from offering several advantages for the way of life, this integration enables a significant amount of routine tasks to be automated Yet, when a gadget is online, it opens it susceptible to hacking attempts by malevolent individuals or other organizations looking to exploit the weaknesses in the device.Growing heterogeneity and diversity of devices increases the frequency of security flaws and increases the difficulty of patching and resolving them.Attacks by hackers that might affect more devices and a larger variety of targets are now more likely to occur.Cybercriminals are using "Distributed Denial-Of-Service"(DDoS) attacks increasingly to undermine systems.This project aims to create a brand-new intrusion detection system powered by deep learning created for the Internet of Things (IoT), since traditional machine learning is not able to detect these threats in realworld deployment.This technique makes the effective claim to identify and neutralize DDoS attacks inside the particular context of networked devices.The proposed hybrid model combines "Recurrent neural networks"(RNN),"long short-term memory" (LSTM), and "Multilayer perceptron"(MLP) to recognize all sorts of DDoS attacks and their specific subcategories.This dataset --CICDDoS-2019--,compiles with everything which satisfies all intrusion detection dataset requirements, is utilized to evaluate the proposed model.