Correlating intrusion detection alerts on bot malware infections using neural network

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

Resumé

Millions of computers are infected with bot malware, form botnets and enable botmaster to perform malicious and criminal activities. Intrusion Detection Systems are deployed to detect infections, but they raise many correlated alerts for each infection, requiring a large manual investigation effort. This paper presents a novel method with a goal of determining which alerts are correlated, by applying Neural Networks and clustering, thus reducing the number of alerts to manually process. The main advantage of the method is that no domain knowledge is required for designing feature extraction or any other part, as such knowledge is inferred by Neural Networks. Evaluation has been performed with traffic traces of real bot binaries executed in a lab setup. The method is trained on labelled Intrusion Detection System alerts and is capable of correctly predicting which of seven incidents an alert pertains, 56.15% of the times. Based on the observed performance it is concluded that the task of understanding Intrusion Detection System alerts can be handled by a Neural Network, showing the potential for reducing the need for manual processing of alerts. Finally, it should be noted that, this is achieved without any feature engineering and with no use of domain specific knowledge.
OriginalsprogEngelsk
TitelCyber Security And Protection Of Digital Services (Cyber Security), 2016 International Conference On
Antal sider8
ForlagIEEE
Publikationsdatojul. 2016
ISBN (Trykt)978-1-5090-0710-3
ISBN (Elektronisk)978-1-5090-0709-7
DOI
StatusUdgivet - jul. 2016
Begivenhed2016 International Conference On Cyber Security And Protection Of Digital Services - Holiday Inn London Mayfair, London, Storbritannien
Varighed: 13 jun. 201614 jun. 2016

Konference

Konference2016 International Conference On Cyber Security And Protection Of Digital Services
LokationHoliday Inn London Mayfair
LandStorbritannien
ByLondon
Periode13/06/201614/06/2016
NavnInternational Conference On Cyber Security And Protection Of Digital Services (Cyber Security). Proceedings.

Fingerprint

Intrusion detection
Neural networks
Feature extraction
Processing
Malware

Emneord

    Citer dette

    Kidmose, E., Stevanovic, M., & Pedersen, J. M. (2016). Correlating intrusion detection alerts on bot malware infections using neural network. I Cyber Security And Protection Of Digital Services (Cyber Security), 2016 International Conference On IEEE. International Conference On Cyber Security And Protection Of Digital Services (Cyber Security). Proceedings. https://doi.org/10.1109/CyberSecPODS.2016.7502344
    Kidmose, Egon ; Stevanovic, Matija ; Pedersen, Jens Myrup. / Correlating intrusion detection alerts on bot malware infections using neural network. Cyber Security And Protection Of Digital Services (Cyber Security), 2016 International Conference On. IEEE, 2016. (International Conference On Cyber Security And Protection Of Digital Services (Cyber Security). Proceedings.).
    @inproceedings{182b07f182f64476ac82699fab41da34,
    title = "Correlating intrusion detection alerts on bot malware infections using neural network",
    abstract = "Millions of computers are infected with bot malware, form botnets and enable botmaster to perform malicious and criminal activities. Intrusion Detection Systems are deployed to detect infections, but they raise many correlated alerts for each infection, requiring a large manual investigation effort. This paper presents a novel method with a goal of determining which alerts are correlated, by applying Neural Networks and clustering, thus reducing the number of alerts to manually process. The main advantage of the method is that no domain knowledge is required for designing feature extraction or any other part, as such knowledge is inferred by Neural Networks. Evaluation has been performed with traffic traces of real bot binaries executed in a lab setup. The method is trained on labelled Intrusion Detection System alerts and is capable of correctly predicting which of seven incidents an alert pertains, 56.15{\%} of the times. Based on the observed performance it is concluded that the task of understanding Intrusion Detection System alerts can be handled by a Neural Network, showing the potential for reducing the need for manual processing of alerts. Finally, it should be noted that, this is achieved without any feature engineering and with no use of domain specific knowledge.",
    keywords = "Artificial neural networks, Clustering algorithms, Correlation, Intrusion detection, Knowledge engineering, Neurons, Training",
    author = "Egon Kidmose and Matija Stevanovic and Pedersen, {Jens Myrup}",
    year = "2016",
    month = "7",
    doi = "10.1109/CyberSecPODS.2016.7502344",
    language = "English",
    isbn = "978-1-5090-0710-3",
    booktitle = "Cyber Security And Protection Of Digital Services (Cyber Security), 2016 International Conference On",
    publisher = "IEEE",
    address = "United States",

    }

    Kidmose, E, Stevanovic, M & Pedersen, JM 2016, Correlating intrusion detection alerts on bot malware infections using neural network. i Cyber Security And Protection Of Digital Services (Cyber Security), 2016 International Conference On. IEEE, International Conference On Cyber Security And Protection Of Digital Services (Cyber Security). Proceedings., London, Storbritannien, 13/06/2016. https://doi.org/10.1109/CyberSecPODS.2016.7502344

    Correlating intrusion detection alerts on bot malware infections using neural network. / Kidmose, Egon; Stevanovic, Matija; Pedersen, Jens Myrup.

    Cyber Security And Protection Of Digital Services (Cyber Security), 2016 International Conference On. IEEE, 2016.

    Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

    TY - GEN

    T1 - Correlating intrusion detection alerts on bot malware infections using neural network

    AU - Kidmose, Egon

    AU - Stevanovic, Matija

    AU - Pedersen, Jens Myrup

    PY - 2016/7

    Y1 - 2016/7

    N2 - Millions of computers are infected with bot malware, form botnets and enable botmaster to perform malicious and criminal activities. Intrusion Detection Systems are deployed to detect infections, but they raise many correlated alerts for each infection, requiring a large manual investigation effort. This paper presents a novel method with a goal of determining which alerts are correlated, by applying Neural Networks and clustering, thus reducing the number of alerts to manually process. The main advantage of the method is that no domain knowledge is required for designing feature extraction or any other part, as such knowledge is inferred by Neural Networks. Evaluation has been performed with traffic traces of real bot binaries executed in a lab setup. The method is trained on labelled Intrusion Detection System alerts and is capable of correctly predicting which of seven incidents an alert pertains, 56.15% of the times. Based on the observed performance it is concluded that the task of understanding Intrusion Detection System alerts can be handled by a Neural Network, showing the potential for reducing the need for manual processing of alerts. Finally, it should be noted that, this is achieved without any feature engineering and with no use of domain specific knowledge.

    AB - Millions of computers are infected with bot malware, form botnets and enable botmaster to perform malicious and criminal activities. Intrusion Detection Systems are deployed to detect infections, but they raise many correlated alerts for each infection, requiring a large manual investigation effort. This paper presents a novel method with a goal of determining which alerts are correlated, by applying Neural Networks and clustering, thus reducing the number of alerts to manually process. The main advantage of the method is that no domain knowledge is required for designing feature extraction or any other part, as such knowledge is inferred by Neural Networks. Evaluation has been performed with traffic traces of real bot binaries executed in a lab setup. The method is trained on labelled Intrusion Detection System alerts and is capable of correctly predicting which of seven incidents an alert pertains, 56.15% of the times. Based on the observed performance it is concluded that the task of understanding Intrusion Detection System alerts can be handled by a Neural Network, showing the potential for reducing the need for manual processing of alerts. Finally, it should be noted that, this is achieved without any feature engineering and with no use of domain specific knowledge.

    KW - Artificial neural networks

    KW - Clustering algorithms

    KW - Correlation

    KW - Intrusion detection

    KW - Knowledge engineering

    KW - Neurons

    KW - Training

    UR - http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=7502344&refinements%3D4225576815%26filter%3DAND%28p_IS_Number%3A7502334%29

    U2 - 10.1109/CyberSecPODS.2016.7502344

    DO - 10.1109/CyberSecPODS.2016.7502344

    M3 - Article in proceeding

    SN - 978-1-5090-0710-3

    BT - Cyber Security And Protection Of Digital Services (Cyber Security), 2016 International Conference On

    PB - IEEE

    ER -

    Kidmose E, Stevanovic M, Pedersen JM. Correlating intrusion detection alerts on bot malware infections using neural network. I Cyber Security And Protection Of Digital Services (Cyber Security), 2016 International Conference On. IEEE. 2016. (International Conference On Cyber Security And Protection Of Digital Services (Cyber Security). Proceedings.). https://doi.org/10.1109/CyberSecPODS.2016.7502344