Distributed Denial of Service Attack Detection for the Internet of Things Using Hybrid Deep Learning Model

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 119862 - 119875

Published: Jan. 1, 2023

As a result of the widespread adoption Internet Things, there are now hundreds millions connected devices, increasing likelihood that they may be vulnerable to various types cyberattacks. In recent years, distributed denial service (DDoS) has emerged as one most destructive tools utilized by attackers. Traditional machine learning approaches typically ineffective and unable cope with actual traffic properties when used identify DDoS attacks. This paper introduces novel deep learning-based intrusion detection system, specifically designed for deployment at either Cloud or Fog level in IoT environment. The proposed model aims detect all attacks their specific subcategory. Our hybrid combines different models, including Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM), Deep Autoencoder, (DNNs). is made up two main levels. first contains parallel sub-neural networks trained algorithms. second uses output frozen combined initial data input. idea behind combination these neural exploit achieve very high performance. To evaluate our model, we CIC-DDoS2019 dataset, which satisfies constraints an dataset. results obtained demonstrate outperformed well-known models terms true positive rate, accuracy, false alarm average rate.

Language: Английский

---

Multilevel Deep Neural Network Approach for Enhanced Distributed Denial-of-Service Attack Detection and Classification in Software-Defined Internet of Things Networks

IEEE Internet of Things Journal, Journal Year: 2024, Volume and Issue: 11(14), P. 24715 - 24725

Published: March 12, 2024

With the increasing rates of interconnected Internet Things (IoT) devices within Software-Defined Networking (SDN) environments, distributed denial service (DDoS) attacks have become increasingly common. As a result this challenge, novel detection and classification methods must be developed based on unique characteristics SDN-supported IoT networks. This paper proposes approach to detecting categorizing DDoS that has been optimized specifically for such environments. part our methodology, we integrate convolutional neural networks (CNN) long-short-term memory (LSTM) models into multilevel deep network architecture. hybrid architecture, complex spatial temporal patterns can automatically extracted from raw traffic data facilitate comprehensive analysis accurate identification attacks. We validate efficacy superiority proposed over traditional machine learning algorithms by conducting rigorous experiments real-world datasets. Our findings underscore potential multi-level as robust scalable solution mitigating in By improving security resilience evolving threats, methodology contributes safeguarding critical infrastructures era ecosystems.

Language: Английский

---

A deep learning technique to detect distributed denial of service attacks in software-defined networks

Computers & Security, Journal Year: 2023, Volume and Issue: 137, P. 103588 - 103588

Published: Nov. 10, 2023

Language: Английский

---

Feature-Selection-Based DDoS Attack Detection Using AI Algorithms

Telecom, Journal Year: 2024, Volume and Issue: 5(2), P. 333 - 346

Published: April 17, 2024

SDN has the ability to transform network design by providing increased versatility and effective regulation. Its programmable centralized controller gives administration employees more authority, allowing for seamless supervision. However, centralization makes it vulnerable a variety of attack vectors, with distributed denial service (DDoS) attacks posing serious concern. Feature selection-based Machine Learning (ML) techniques are than traditional signature-based Intrusion Detection Systems (IDS) at identifying new threats in context defending against attacks. In this study, NGBoost is compared four additional machine learning algorithms: convolutional neural (CNN), Stochastic Gradient Descent (SGD), Decision Tree, Random Forest, order assess effectiveness DDoS detection on CICDDoS2019 dataset. It focuses important measures such as F1 score, recall, accuracy, precision. We have examined NeTBIOS, layer-7 attack, SYN, layer-4 our paper. Our investigation shows that Natural Boosting Convolutional Neural Networks, particular, show promise tabular data categorization. conclusion, we go through specific study results protecting using DDoS. These experimental findings offer framework making decisions.

Language: Английский

---

Smart network security using advanced ensemble-DDoS attack detection and hybrid JA-SLOA-linked optimal routing-based mitigation

Australian Journal of Electrical & Electronics Engineering, Journal Year: 2024, Volume and Issue: 21(4), P. 374 - 396

Published: April 21, 2024

Distributed Denial of Service (DDoS) attacks are distributed at a faster rate, and they considered to be fatal threats over the Internet. Moreover, several deep learning approaches insufficient attain maximum efficiency appropriate detection due complexity diversity DDoS attack traffic under fast fast-speed network environment since providing with individual performance. Hence, an ensemble model is developed for mitigation hybrid optimization algorithm ensure good detective performance against attacks. The pre-processed data fed feature extraction process where it done through Deep Belief Network (DBN) Autoencoder techniques acquiring features. optimized fusion features takes place using meta-heuristic Adaptive Sound Speed-based Jaya Sea Lion Optimization (ASS-JSLnO). performed by Improved Ensemble Learning (IEL) approach. Then, strategy applied optimal routing multi-objective function. Finally, experiments made establish high efficacy implemented framework.

Language: Английский

---

Lightweight Deep Learning Method based on Group Convolution: Detecting DDoS Attacks in IoT Environments

Published: March 16, 2024

Distributed denial-of-service (DDoS) attacks remain one of the major security threats in Internet Things (IoT) domain. Compared to traditional computing devices, IoT devices typically have more limited computational capabilities and memory resources. To address resource constraints DDoS detection, this study proposes a lightweight detection model called DGConv-IDS based on autoencoders convolutional neural networks. adopts sliding window algorithm only retain recent data, effectively controlling overhead by leveraging temporal features traffic. The integrates dynamic group networks into unified framework, where are used for unsupervised feature extraction dimensionality reduction, graph modules real-time different types attacks. For publicly available datasets such as CICIoT2023, we extract multi-dimensional including timestamps, IP addresses, packet sizes, protocol types, etc. train model. Experimental results show that achieves high accuracy multiple datasets. with similar deep learning-based methods, has lower costs better performance. In general, proposed is expected improve protection provide effective solutions help systems resist

Language: Английский

---

Performance evaluation of machine learning models for distributed denial of service attack detection using improved feature selection and hyper‐parameter optimization techniques

Concurrency and Computation Practice and Experience, Journal Year: 2022, Volume and Issue: 34(26)

Published: Aug. 29, 2022

Summary This article gives the framework of extensive experimentation various machine learning models to detect distributed denial service attacks (DDoS). We use six‐tier feature ranking methods that statistical techniques as well based classifiers obtain significant features. The measurable selection involves Chi‐Square (Chi2), information gain (IG), merged (Chi2)‐IG and involve ensemble classifiers, is, decision tree, random forest eXtreme gradient boosting (XGBoost). Different supervised (logistic regression, tree classifier, linear support vector machine, k‐nearest neighbors, Gaussian Naive Bayes, XGBoost) are trained on a feature‐engineered datasets. To further our research, we neural networks (ANN CNN) using both feature‐selected auto‐feature training setup. check validation adaptability these with optimal tuning parameters GridSearchCV effectiveness sampling in overcoming class imbalance problem. Based methods, evaluated for their best performance. experimental results show outperformed ones state art. performance analysis is done confusion matrix scores, accuracy, false alarm rate, sensitivity, specificity, false‐positive F1 score, area under curve loss functions well‐known KDD Cup 99 UNSW‐NB15 study furthering research DDoS detection deep networks.

Language: Английский

---

A Wrapper Feature Selection Based Hybrid Deep Learning Model for DDoS Detection in a Network with NFV Behaviors

Wireless Personal Communications, Journal Year: 2023, Volume and Issue: 133(1), P. 481 - 506

Published: Nov. 1, 2023

Language: Английский

---

A DoS attack detection method based on adversarial neural network

PeerJ Computer Science, Journal Year: 2024, Volume and Issue: 10, P. e2162 - e2162

Published: Aug. 9, 2024

In order to analyze the influence of deep learning model on detecting denial-of-service (DoS) attacks, this article first examines concepts and attack strategies DoS assaults before looking into present detection methodologies for attacks. A distributed system based is established in response investigation’s limitations. This can quickly accurately identify traffic attacks network that needs be detected then promptly send an alarm signal system. Then, a called Improved Conditional Wasserstein Generative Adversarial Network with Inverter (ICWGANInverter) proposed characteristics incomplete automatically learns advanced abstract information original data employs method reconstruction error best classification label. It tested intrusion dataset NSL-KDD. The findings demonstrate mean square continuous feature sub-datasets KDDTest+ KDDTest-21 steadily increases as noise factor increases. All receiver operating characteristic (ROC) curves are shown at top diagonal, overall area under ROC curve (AUC) values macro-average micro-average above 0.8, which demonstrates ICWGANInverter has excellent performance both single category detection. greater accuracy than other models, reaching 87.79%. approach suggested offers higher benefits

Language: Английский

---

Research on the Characteristics and Detection Methods of DDoS Attacks on Wireless Sensor Networks for Vehicle Networking

Engineering Advances, Journal Year: 2023, Volume and Issue: 2(2), P. 175 - 181

Published: Jan. 12, 2023

Smart vehicles constitute the intelligent transportation system, complex traffic network of multiple types sensors in energy consumption data and amount transmitted is increasing, consisting multi-source wireless vehicle often subject to DDoS attacks, will lead loss or even failure.Since distributed nodes are dynamic constantly entering leaving a cluster, as smart continue join new sensor obtain identity IDs based on location, IP addresses always allocated recycled.DDoS attacks against networking clusters difficult identify, destructive easy implement.In this paper, we analyze topology communication patterns networks vehicular networks, characteristics being detection methods each, propose initial trust value calculation for attack nodes.

Language: Английский

---