Roberto Doriguzzi-Corin

Maged Abdelaty, Sandra Scott-Hayward, Roberto Doriguzzi-Corin et al.

Machine Learning (ML) has proven to be effective in many application domains. However, ML methods can be vulnerable to adversarial attacks, in which an attacker tries to fool the classification/prediction mechanism by crafting the input data. In the case of ML-based Network Intrusion Detection Systems (NIDSs), the attacker might use their knowledge of the intrusion detection logic to generate malicious traffic that remains undetected. One way to solve this issue is to adopt adversarial training, in which the training set is augmented with adversarial traffic samples. This paper presents an adversarial training approach called GADoT, which leverages a Generative Adversarial Network (GAN) to generate adversarial DDoS samples for training. We show that a state-of-the-art NIDS with high accuracy on popular datasets can experience more than 60% undetected malicious flows under adversarial attacks. We then demonstrate how this score drops to 1.8% or less after adversarial training using GADoT.

Maged Abdelaty, Roberto Doriguzzi-Corin, Domenico Siracusa

Deep Learning is emerging as an effective technique to detect sophisticated cyber-attacks targeting Industrial Control Systems (ICSs). The conventional approach to detection in literature is to learn the "normal" behaviour of the system, to be then able to label noteworthy deviations from it as anomalies. However, during operations, ICSs inevitably and continuously evolve their behaviour, due to e.g., replacement of devices, workflow modifications, or other reasons. As a consequence, the accuracy of the anomaly detection process may be dramatically affected with a considerable amount of false alarms being generated. This paper presents DAICS, a novel deep learning framework with a modular design to fit in large ICSs. The key component of the framework is a 2-branch neural network that learns the changes in the ICS behaviour with a small number of data samples and a few gradient updates. This is supported by an automatic tuning mechanism of the detection threshold that takes into account the changes in the prediction error under normal operating conditions. In this regard, no specialised human intervention is needed to update the other parameters of the system. DAICS has been evaluated using publicly available datasets and shows an increased detection rate and accuracy compared to state of the art approaches, as well as higher robustness to additive noise.

Roberto Doriguzzi-Corin

With the recent trend of "network softwarisation", enabled by emerging technologies such as Software-Defined Networking (SDN) and Network Function Virtualisation (NFV), system administrators of data centres and enterprise networks have started replacing dedicated hardware-based middleboxes with virtualised network functions running on servers and end hosts. This radical change has facilitated the provisioning of advanced and flexible network services, ultimately helping system administrators and network operators to cope with the rapid changes in service requirements and networking workloads. This thesis investigates the challenges of provisioning network security services in "softwarised" networks, where the security of residential and business users can be provided by means of sets of software-based network functions running on high performance servers or on commodity compute devices. The study is approached from the perspective of the telecom operator, whose goal is to protect the customers from network threats and, at the same time, maximize the number of provisioned services, and thereby revenue. Specifically, the overall aim of the research presented in this thesis is proposing novel techniques for optimising the resource usage of software-based security services, hence for increasing the chances for the operator to accommodate more service requests while respecting the desired level of network security of its customers. In this direction, the contributions of this thesis are the following: (i) a solution for the dynamic provisioning of security services that minimises the utilisation of computing and network resources, and (ii) novel methods based on Deep Learning and Linux kernel technologies for reducing the CPU usage of software-based security network functions, with specific focus on the defence against Distributed Denial of Service (DDoS) attacks.

Roberto Doriguzzi-Corin, Stuart Millar, Sandra Scott-Hayward et al.

Distributed Denial of Service (DDoS) attacks are one of the most harmful threats in today's Internet, disrupting the availability of essential services. The challenge of DDoS detection is the combination of attack approaches coupled with the volume of live traffic to be analysed. In this paper, we present a practical, lightweight deep learning DDoS detection system called LUCID, which exploits the properties of Convolutional Neural Networks (CNNs) to classify traffic flows as either malicious or benign. We make four main contributions; (1) an innovative application of a CNN to detect DDoS traffic with low processing overhead, (2) a dataset-agnostic preprocessing mechanism to produce traffic observations for online attack detection, (3) an activation analysis to explain LUCID's DDoS classification, and (4) an empirical validation of the solution on a resource-constrained hardware platform. Using the latest datasets, LUCID matches existing state-of-the-art detection accuracy whilst presenting a 40x reduction in processing time, as compared to the state-of-the-art. With our evaluation results, we prove that the proposed approach is suitable for effective DDoS detection in resource-constrained operational environments.