Biblio
To assure cyber security of an enterprise, typically SIEM (Security Information and Event Management) system is in place to normalize security events from different preventive technologies and flag alerts. Analysts in the security operation center (SOC) investigate the alerts to decide if it is truly malicious or not. However, generally the number of alerts is overwhelming with majority of them being false positive and exceeding the SOC's capacity to handle all alerts. Because of this, potential malicious attacks and compromised hosts may be missed. Machine learning is a viable approach to reduce the false positive rate and improve the productivity of SOC analysts. In this paper, we develop a user-centric machine learning framework for the cyber security operation center in real enterprise environment. We discuss the typical data sources in SOC, their work flow, and how to leverage and process these data sets to build an effective machine learning system. The paper is targeted towards two groups of readers. The first group is data scientists or machine learning researchers who do not have cyber security domain knowledge but want to build machine learning systems for security operations center. The second group of audiences are those cyber security practitioners who have deep knowledge and expertise in cyber security, but do not have machine learning experiences and wish to build one by themselves. Throughout the paper, we use the system we built in the Symantec SOC production environment as an example to demonstrate the complete steps from data collection, label creation, feature engineering, machine learning algorithm selection, model performance evaluations, to risk score generation.
Used by both information systems designers and security personnel, the Attack Tree method provides a graphical analysis of the ways in which an entity (a computer system or network, an entire organization, etc.) can be attacked and indicates the countermeasures that can be taken to prevent the attackers to reach their objective. In this paper, we built an Attack Tree focused on the goal “compromising the security of a Web platform”, considering the most common vulnerabilities of the WordPress platform identified by CVE (Common Vulnerabilities and Exposures), a global reference system for recording information regarding computer security threats. Finally, based on the likelihood of the attacks, we made a quantitative analysis of the probability that the security of the Web platform can be compromised.
Deep learning techniques have demonstrated the ability to perform a variety of object recognition tasks using visible imager data; however, deep learning has not been implemented as a means to autonomously detect and assess targets of interest in a physical security system. We demonstrate the use of transfer learning on a convolutional neural network (CNN) to significantly reduce training time while keeping detection accuracy of physical security relevant targets high. Unlike many detection algorithms employed by video analytics within physical security systems, this method does not rely on temporal data to construct a background scene; targets of interest can halt motion indefinitely and still be detected by the implemented CNN. A key advantage of using deep learning is the ability for a network to improve over time. Periodic retraining can lead to better detection and higher confidence rates. We investigate training data size versus CNN test accuracy using physical security video data. Due to the large number of visible imagers, significant volume of data collected daily, and currently deployed human in the loop ground truth data, physical security systems present a unique environment that is well suited for analysis via CNNs. This could lead to the creation of algorithmic element that reduces human burden and decreases human analyzed nuisance alarms.
New and unseen network attacks pose a great threat to the signature-based detection systems. Consequently, machine learning-based approaches are designed to detect attacks, which rely on features extracted from network data. The problem is caused by different distribution of features in the training and testing datasets, which affects the performance of the learned models. Moreover, generating labeled datasets is very time-consuming and expensive, which undercuts the effectiveness of supervised learning approaches. In this paper, we propose using transfer learning to detect previously unseen attacks. The main idea is to learn the optimized representation to be invariant to the changes of attack behaviors from labeled training sets and non-labeled testing sets, which contain different types of attacks and feed the representation to a supervised classifier. To the best of our knowledge, this is the first effort to use a feature-based transfer learning technique to detect unseen variants of network attacks. Furthermore, this technique can be used with any common base classifier. We evaluated the technique on publicly available datasets, and the results demonstrate the effectiveness of transfer learning to detect new network attacks.
We present a testbed implementation for the development, evaluation and demonstration of security orchestration in a network function virtualization environment. As a specific scenario, we demonstrate how an intelligent response to DDoS and various other kinds of targeted attacks can be formulated such that these attacks and future variations can be mitigated. We utilise machine learning to characterise normal network traffic, attacks and responses, then utilise this information to orchestrate virtualized network functions around affected components to isolate these components and to capture, redirect and filter traffic (e.g. honeypotting) for additional analysis. This allows us to maintain a high level of network quality of service to given network functions and components despite adverse network conditions.
Supervisory control and data acquisition (SCADA) systems are the key driver for critical infrastructures and industrial facilities. Cyber-attacks to SCADA networks may cause equipment damage or even fatalities. Identifying risks in SCADA networks is critical to ensuring the normal operation of these industrial systems. In this paper we propose a Bayesian network-based cyber-security risk assessment model to dynamically and quantitatively assess the security risk level in SCADA networks. The major distinction of our work is that the proposed risk assessment method can learn model parameters from historical data and then improve assessment accuracy by incrementally learning from online observations. Furthermore, our method is able to assess the risk caused by unknown attacks. The simulation results demonstrate that the proposed approach is effective for SCADA security risk assessment.
Cyber attacks, (e.g., DDoS), on computers connected to the Internet occur everyday. A DDoS attack in 2016 that used “Mirai botnet” generated over 600 Gbit/s traffic, which was twice as that of last year. In view of this situation, we can no longer adequately protect our computers using current end-point security solutions and must therefore introduce a new method of protection that uses distributed nodes, e.g., routers. We propose an Autonomous and Distributed Internet Security (AIS) infrastructure that provides two key functions: first, filtering source address spoofing packets (proactive filter), and second, filtering malicious packets that are observed at the end point (reactive filter) at the closest malicious packets origins. We also propose three types of Multi-Layer Binding Routers (MLBRs) to realize these functions. We implemented the MLBRs and constructed experimental systems to simulate DDoS attacks. Results showed that all malicious packets could be filtered by using the AIS infrastructure.
Emerging nonvolatile memory (NVM) devices are not limited to build nonvolatile memory macros. They can also be used in developing nonvolatile logics (nvLogics) for nonvolatile processors, security circuits for the internet of things (IoT), and computing-in-memory (CIM) for artificial intelligence (AI) chips. This paper explores the challenges in circuit designs of emerging memory devices for application in nonvolatile logics, security circuits, and CIM for deep neural networks (DNN). Several silicon-verified examples of these circuits are reviewed in this paper.
The Air Force is shifting its cybersecurity paradigm from an information technology (IT)-centric toward a mission oriented approach. Instead of focusing on how to defend its IT infrastructure, it seeks to provide mission assurance by defending mission relevant cyber terrain enabling mission execution in a contested environment. In order to actively defend a mission in cyberspace, efforts must be taken to understand and document that mission's dependence on cyberspace and cyber assets. This is known as cyber terrain mission mapping. This paper seeks to define mission mapping and overview methodologies. We also analyze current tools seeking to provide cyber situational awareness through mission mapping or cyber dependency impact analysis and identify existing shortfalls.
Kings Eye is a platform independent situational awareness prototype for smart devices. Platform independence is important as there are more and more soldiers bringing their own devices, with different operating systems, into the field. The concept of Bring Your Own Device (BYOD) is a low-cost approach to equipping soldiers with situational awareness tools and by this it is important to facilitate and evaluate such solutions.
Situational awareness during sophisticated cyber attacks on the power grid is critical for the system operator to perform suitable attack response and recovery functions to ensure grid reliability. The overall theme of this paper is to identify existing practical issues and challenges that utilities face while monitoring substations, and to suggest potential approaches to enhance the situational awareness for the grid operators. In this paper, we provide a broad discussion about the various gaps that exist in the utility industry today in monitoring substations, and how those gaps could be addressed by identifying the various data sources and monitoring tools to improve situational awareness. The paper also briefly describes the advantages of contextualizing and correlating substation monitoring alerts using expert systems at the control center to obtain a holistic systems-level view of potentially malicious cyber activity at the substations before they cause impacts to grid operation.
As societies are becoming more dependent on the power grids, the security issues and blackout threats are more emphasized. This paper proposes a new graph model for online visualization and assessment of power grid security. The proposed model integrates topology and power flow information to estimate and visualize interdependencies between the lines in the form of line dependency graph (LDG) and immediate threats graph (ITG). These models enable the system operator to predict the impact of line outage and identify the most vulnerable and critical links in the power system. Line Vulnerability Index (LVI) and Line Criticality Index (LCI) are introduced as two indices extracted from LDG to aid the operator in decision making and contingency selection. This package can be useful in enhancing situational awareness in power grid operation by visualization and estimation of system threats. The proposed approach is tested for security analysis of IEEE 30-bus and IEEE 118-bus systems and the results are discussed.
This paper introduces SONA (Spatiotemporal system Organized for Natural Analysis), a tabletop and tangible controller system for exploring geotagged information, and more specifically, CCTV. SONA's goal is to support a more natural method of interacting with data. Our new interactions are placed in the context of a physical security environment, closed circuit television (CCTV). We present a three-layered detail on demand set of view filters for CCTV feeds on a digital map. These filters are controlled with a novel tangible device for direct interaction. We validate SONA's tangible controller approach with a user study comparing SONA with the existing CCTV multi-screen method. The results of the study show that SONA's tangible interaction method is superior to the multi-screen approach, both in terms of quantitative results, and is preferred by users.
Cybersecurity is one of critical issues in modern military operations. In cyber operations, security professionals depend on various information and security systems to mitigate cyber threats through enhanced cyber situational awareness. Cyber situational awareness can give decision makers mission completeness and providing appropriate timely decision support for proactive response. The crucial information for cyber situational awareness can be collected at network boundaries through deep packet inspection with security systems. Regular expression is regarded as a practical method for deep packet inspection that is considering a next generation intrusion detection and prevention, however, it is not commonly used by the reason of its resource intensive characteristics. In this paper, we describe our effort and achievement on regular expression processing capability in real time and an evaluation method with experimental result.
Data from cyber logs can often be represented as a bipartite graph (e.g. internal IP-external IP, user-application, or client-server). State-of-the-art graph based anomaly detection often generalizes across all types of graphs — namely bipartite and non-bipartite. This confounds the interpretation and use of specific graph features such as degree, page rank, and eigencentrality that can provide a security analyst with rapid situational awareness of their network. Furthermore, graph algorithms applied to data collected from large, distributed enterprise scale networks require accompanying methods that allow them to scale to the data collected. In this paper, we provide a novel, scalable, directional graph projection framework that operates on cyber logs that can be represented as bipartite graphs. This framework computes directional graph projections and identifies a set of interpretable graph features that describe anomalies within each partite.
As one of the security components in cyber situational awareness systems, Intrusion Detection System (IDS) is implemented by many organizations in their networks to address the impact of network attacks. Regardless of the tools and technologies used to generate security alarms, IDS can provide a situation overview of network traffic. With the security alarm data generated, most organizations do not have the right techniques and further analysis to make this alarm data more valuable for the security team to handle attacks and reduce risk to the organization. This paper proposes the IDS Metrics Framework for cyber situational awareness system that includes the latest technologies and techniques that can be used to create valuable metrics for security advisors in making the right decisions. This metrics framework consists of the various tools and techniques used to evaluate the data. The evaluation of the data is then used as a measurement against one or more reference points to produce an outcome that can be very useful for the decision making process of cyber situational awareness system. This metric offers an additional Graphical User Interface (GUI) tools that produces graphical displays and provides a great platform for analysis and decision-making by security teams.
Cyber Physical Systems (CPS) operating in modern critical infrastructures (CIs) are increasingly being targeted by highly sophisticated cyber attacks. Threat actors have quickly learned of the value and potential impact of targeting CPS, and numerous tailored multi-stage cyber-physical attack campaigns, such as Advanced Persistent Threats (APTs), have been perpetrated in the last years. They aim at stealthily compromising systems' operations and cause severe impact on daily business operations such as shutdowns, equipment damage, reputation damage, financial loss, intellectual property theft, and health and safety risks. Protecting CIs against such threats has become as crucial as complicated. Novel distributed detection and reaction methodologies are necessary to effectively uncover these attacks, and timely mitigate their effects. Correlating large amounts of data, collected from a multitude of relevant sources, is fundamental for Security Operation Centers (SOCs) to establish cyber situational awareness, and allow to promptly adopt suitable countermeasures in case of attacks. In our previous work we introduced three methods for security information correlation. In this paper we define metrics and benchmarks to evaluate these correlation methods, we assess their accuracy, and we compare their performance. We finally demonstrate how the presented techniques, implemented within our cyber threat intelligence analysis engine called CAESAIR, can be applied to support incident handling tasks performed by SOCs.