Generative Adversarial Networks (GANs): A Survey on Network Traffic Generation

doi:10.18178/ijmlc.2022.12.6.1120

Cited by 4 publications

(2 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed methods differ from traditional protocol-based data generation/classification, which relies on rules, heuristics, or information about IP, port numbers, and signature protocols. Several methods, including Bayesian methods (specifically neural networks), modified Association algorithms, Support Vector Machines, Venn Probability Machines, k-Nearest Neighbor, and k-means clustering algorithms, have been employed to produce class probability distributions [22]. Also, neural networks, particularly deep neural networks, have demonstrated tremendous potential in the area of generation [21], [22].…”

Section: A Chronological Survey Of Network Traffic Generatorsmentioning

confidence: 99%

“…GAN-based architectures have become capable of comprehending intricate data distributions at the packet level and producing comparable outcomes with minor differences throughout the recent years. This has been achieved through certain modifications made to the model, as described in [21], [22]. GAN implementations commonly generate data in three different formats: packet, flow, or tabular.…”

Section: Generative Adversarial Networkmentioning

confidence: 99%

See 1 more Smart Citation

PAC-GPT: A Novel Approach to Generating Synthetic Network Traffic With GPT-3

Kholgh,

Kostakos

2023

IEEE Access

View full text Add to dashboard Cite

The application of machine learning models, particularly in cybersecurity, has surged significantly in the past few years. However, the effectiveness of these models is predominantly tethered to the quality and breadth of the training data they ingest. The scarcity of realistic datasets within the cybersecurity field constitutes a considerable challenge to the development of industry-grade tools intended for real-world application scenarios. Specifically, current datasets are either significantly outdated or fall short on both qualitative and quantitative fronts, primarily because many organizations exhibit reluctance in data sharing, stemming from privacy concerns or the potential threat to trade secrets. To address this challenge, the paper introduces PAC-GPT, a novel framework to generate reliable synthetic data for machine learning methods based on Open AI's Generative Pre-trained Transformer 3 (GPT-3). The core components of this framework are two modules, namely a Flow Generator, which is responsible for capturing and regenerating patterns in a series of network packets, and Packet Generator, which can generate individual network packets given the network flow. We also propose a packet generator based on LLM chaining and then proceed to assess, compare, and evaluate its performance using metrics such as loss, accuracy and success rate, concluding that transformers are a suitable approach for synthetic packet generation with minimal fine-tuning performed. Lastly, a streamlined command line interface (CLI) tool has been devised to facilitate the seamless access of this innovative data generation strategy by professionals from various disciplines.

show abstract

Section: A Chronological Survey Of Network Traffic Generatorsmentioning

confidence: 99%

Section: Generative Adversarial Networkmentioning

confidence: 99%

PAC-GPT: A Novel Approach to Generating Synthetic Network Traffic With GPT-3

Kholgh,

Kostakos

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

Machine Learning With Computer Networks: Techniques, Datasets, and Models

Afifi,

Pochaba,

Boltres

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

GothX: a generator of customizable, legitimate and malicious IoT network traffic

Poisson,

Carnier,

Fukuda

2024

Proceedings of the 17th Cyber Security Experimentation and Test Workshop

View full text Add to dashboard Cite

In recent years, machine learning-based anomaly detection (AD) has become an important measure against security threats from Internet of Things (IoT) networks. Machine learning (ML) models for network traffic AD require datasets to be trained, evaluated and compared. Due to the necessity of realistic and up-to-date representation of IoT security threats, new datasets need to be constantly generated to train relevant AD models. Since most traffic generation setups are developed considering only the author's use, replication of traffic generation becomes an additional challenge to the creation and maintenance of useful datasets. In this work, we propose GothX, a flexible traffic generator to create both legitimate and malicious traffic for IoT datasets. As a fork of Gotham Testbed, GothX is developed with five requirements: 1) easy configuration of network topology, 2) customization of traffic parameters, 3) automatic execution of legitimate and attack scenarios, 4) IoT network heterogeneity (the current iteration supports MQTT, Kafka and SINETStream services), and 5) automatic labeling of generated datasets. GothX is validated by two use cases: a) re-generation and enrichment of traffic from the IoT dataset MQTTset,and b) automatic execution of a new realistic scenario including the exploitation of a CVE specific to the Kafka-MQTT network topology and leading to a DDoS attack. We also contribute with two datasets containing mixed traffic, one made from the enriched MQTTset traffic and another from the attack scenario. We evaluated the scalability of GothX (450 IoT sensors in a single machine), the replication of the use cases and the validity of the generated datasets, confirming the ability of GothX to improve the current state-of-the-art of network traffic generation.

show abstract

Generative Adversarial Networks (GANs): A Survey on Network Traffic Generation

Cited by 4 publications

References 40 publications

PAC-GPT: A Novel Approach to Generating Synthetic Network Traffic With GPT-3

PAC-GPT: A Novel Approach to Generating Synthetic Network Traffic With GPT-3

Machine Learning With Computer Networks: Techniques, Datasets, and Models

GothX: a generator of customizable, legitimate and malicious IoT network traffic

Contact Info

Product

Resources

About