Application of Machine Learning-Based Classifier for AS-Sets to Enhance the Security within the Border Gateway Protocol
Hasanov, Ismayil (2023-08-01)
Application of Machine Learning-Based Classifier for AS-Sets to Enhance the Security within the Border Gateway Protocol
Hasanov, Ismayil
(01.08.2023)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
suljettu
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe20230912123908
https://urn.fi/URN:NBN:fi-fe20230912123908
Tiivistelmä
The internet has precipitated a revolutionary transformation across multiple dimensions of human existence, encompassing communication, work, and education. With its intricate mechanisms, the internet relies on the functioning of the Border Gateway Protocol (BGP) to facilitate the exchange of routing information between Internet Service Providers (ISPs) and organizations. However, BGP is exposed to numerous security vulnerabilities that can be exploited by malicious actors, leading to potentially severe and adverse consequences.
This thesis endeavors to investigate the viability of leveraging Artificial Intelligence (AI) models as a solution to bolster BGP security. Specifically, the focus lies on the development of an AI model capable of analyzing BGP AS-Sets and effectively classifying them as either legitimate or suspicious. Additionally, this thesis conducts an in-depth analysis of the role that Large Language Models (LLMs) could play in the field of cybersecurity. To achieve this, experimental work is conducted, leveraging LLMs to develop the necessary code to fulfill the primary objectives.
The structure of the thesis encompasses a comprehensive literature review on the BGP protocol, shedding light on its intricacies and associated vulnerabilities. Furthermore, a literature review on AI and the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework is conducted. This provides a solid foundation for the subsequent practical implementation of the AI-based classifier. The development and testing of the classifier are carried out in collaboration with Openfactory Nordics Oy.
The thesis presents a proficient machine learning-based model, devised to adeptly distinguish between suspicious and legitimate AS-Set instances. This enhances the overall security of BGP.
In conclusion, the thesis presents a summary of the findings and conclusions derived from the extensive research conducted. Moreover, the thesis identifies and explores potential avenues for future research, thereby paving the way for further advancements in the field.
This thesis endeavors to investigate the viability of leveraging Artificial Intelligence (AI) models as a solution to bolster BGP security. Specifically, the focus lies on the development of an AI model capable of analyzing BGP AS-Sets and effectively classifying them as either legitimate or suspicious. Additionally, this thesis conducts an in-depth analysis of the role that Large Language Models (LLMs) could play in the field of cybersecurity. To achieve this, experimental work is conducted, leveraging LLMs to develop the necessary code to fulfill the primary objectives.
The structure of the thesis encompasses a comprehensive literature review on the BGP protocol, shedding light on its intricacies and associated vulnerabilities. Furthermore, a literature review on AI and the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework is conducted. This provides a solid foundation for the subsequent practical implementation of the AI-based classifier. The development and testing of the classifier are carried out in collaboration with Openfactory Nordics Oy.
The thesis presents a proficient machine learning-based model, devised to adeptly distinguish between suspicious and legitimate AS-Set instances. This enhances the overall security of BGP.
In conclusion, the thesis presents a summary of the findings and conclusions derived from the extensive research conducted. Moreover, the thesis identifies and explores potential avenues for future research, thereby paving the way for further advancements in the field.