As cyberattacks become more intelligent, it is challenging to detect advanced attacks in a variety of fields including industry, national defense, and healthcare. Traditional intrusion detection systems are no longer enough to detect these advanced attacks with unexpected patterns. Attackers bypass known signatures and pretend to be normal users. Deep learning is an alternative to solving these issues. Deep Learning (DL)-based intrusion detection does not require a lot of attack signatures or the list of normal behaviors to generate detection rules. DL defines intrusion features by itself through training empirical data. We develop a DL-based intrusion model especially focusing on denial of service (DoS) attacks. For the intrusion dataset, we use KDD CUP 1999 dataset (KDD), the most widely used dataset for the evaluation of intrusion detection systems (IDS). KDD consists of four types of attack categories, such as DoS, user to root (U2R), remote to local (R2L), and probing. Numerous KDD studies have been employing machine learning and classifying the dataset into the four categories or into two categories such as attack and benign. Rather than focusing on the broad categories, we focus on various attacks belonging to same category. Unlike other categories of KDD, the DoS category has enough samples for training each attack. In addition to KDD, we use CSE-CIC-IDS2018 which is the most up-to-date IDS dataset. CSE-CIC-IDS2018 consists of more advanced DoS attacks than that of KDD. In this work, we focus on the DoS category of both datasets and develop a DL model for DoS detection. We develop our model based on a Convolutional Neural Network (CNN) and evaluate its performance through comparison with an Recurrent Neural Network (RNN). Furthermore, we suggest the optimal CNN design for the better performance through numerous experiments.

As the number of Internet of Things (IoT) devices connected to the network rapidly increases, network attacks such as flooding and Denial of Service (DoS) are also increasing. These attacks cause network disruption and denial of service to IoT devices. However, a large number of heterogenous devices deployed in the IoT environment make it difficult to detect IoT attacks using traditional rule-based security solutions. It is challenging to develop optimal security models for each type of the device. Machine learning (ML) is an alternative technique that allows one to develop optimal security models based on empirical data from each device. We employ the ML technique for IoT attack detection. We focus on botnet attacks targeting various IoT devices and develop ML-based models for each type of device. We use the N-BaIoT dataset generated by injecting botnet attacks (Bashlite and Mirai) into various types of IoT devices, including a Doorbell, Baby Monitor, Security Camera, and Webcam. We develop a botnet detection model for each device using numerous ML models, including deep learning (DL) models. We then analyze the effective models with a high detection F1-score by carrying out multiclass classification, as well as binary classification, for each model.

As attack techniques become more sophisticated, detecting new and advanced cyberattacks with traditional intrusion detection techniques based on signature and anomaly is becoming challenging. In signature-based detection, not only do attackers bypass known signatures, but they also exploit unknown vulnerabilities. As the number of new signatures is increasing daily, it is also challenging to scale the detection mechanisms without impacting performance. For anomaly detection, defining normal behaviors is challenging due to today’s complex applications with dynamic features. These complex and dynamic characteristics cause much false positives with a simple outlier detection. In this work, we detect intrusion behaviors by looking at number of computing elements together in time and space, whereas most of existing intrusion detection systems focus on a single element. In order to define the spatiotemporal intrusion patterns, we look at fundamental behaviors of cyberattacks that should appear in any possible attacks. We define these individual behaviors as basic cyberattack action (BCA) and develop a stochastic graph model to represent combination of BCAs in time and space. In addition, we build an intrusion detection system to demonstrate the detection mechanism based on the graph model. We inject numerous known and possible unknown attacks comprising BCAs and show how the system detects these attacks and how to locate the root causes based on the spatiotemporal patterns. The characterization of attacks in spatiotemporal patterns with expected essential behaviors would present a new effective approach to the intrusion detection.

Security administrators of companies and organizations need to come up with proper countermeasures against cyber-attacks considering infrastructures and security policies in their possession. In order to develop and verify such countermeasures, the administrators should be able to reenact both cyber-attacks and defenses. Simulations can be useful for the reenactment by overcoming its limitations including high risk and cost. If the administrators are able to design various scenarios of cyber-attacks and to develop simulation models from their viewpoints, they can simulate desired situations and observe the results more easily. It is challenging to simulate cyber-security issues, because there is lack of theoretical basis for modeling a wide range of the security field as well as pre-defined basic components used to model cyber-attacks. In this paper, we propose a modeling method for cyber-security simulations by developing a basic component and a composite model, called Abstracted Cyber-Security Unit Model (ACSUM) and Abstracted Cyber-security SIMulation model (ACSIM), respectively. The proposed models are based on DEVS(Discrete Event systems Specification) formalism, a modeling theory for discrete event simulations. We develop attack scenarios by sequencing attack behaviors using ACSUMs and then model ACSIMs by combining and abstracting the ACSUMs from a security perspective. The concepts of ACSUM and ACSIM enable the security administrators to simulate numerous cyber-security issues from their viewpoints. As a case study, we model a worm scenario using ACSUM and simulate three types of simulation models based on ACSIM from a different security perspective.