Biblio

With the increasing scale of network, attacks against network emerge one after another, and security problems become increasingly prominent. Network intrusion detection system is a widely used and effective security means at present. In addition, with the development of machine learning technology, various intelligent intrusion detection algorithms also start to sprout. By flexibly combining these intelligent methods with intrusion detection technology, the comprehensive performance of intrusion detection can be improved, but the vulnerability of machine learning model in the adversarial environment can not be ignored. In this paper, we study the defense problem of network intrusion detection system against adversarial samples. More specifically, we design a defense algorithm for NIDS against adversarial samples by using bidirectional generative adversarial network. The generator learns the data distribution of normal samples during training, which is an implicit model reflecting the normal data distribution. After training, the adversarial sample detection module calculates the reconstruction error and the discriminator matching error of sample. Then, the adversarial samples are removed, which improves the robustness and accuracy of NIDS in the adversarial environment.

Intrusion Detection systems are important tools in preventing malicious traffic from penetrating into networks and systems. Recently, Intrusion Detection Systems are rapidly enhancing their detection capabilities using machine learning algorithms. However, these algorithms are vulnerable to new unknown types of attacks that can evade machine learning IDS. In particular, they may be vulnerable to attacks based on Generative Adversarial Networks (GAN). GANs have been widely used in domains such as image processing, natural language processing to generate adversarial data of different types such as graphics, videos, texts, etc. We propose a model using GAN to generate adversarial DDoS attacks that can change the attack profile and can be undetected. Our simulation results indicate that by continuous changing of attack profile, defensive systems that use incremental learning will still be vulnerable to new attacks.

Software Defined Networking (SDN) focus on the isolation of control plane and data plane, greatly enhancing the network's support for heterogeneity and flexibility. However, although the programmable network greatly improves the performance of all aspects of the network, flexible load balancing across controllers still challenges the current SDN architecture. Complex application scenarios lead to flexible and changeable communication requirements, making it difficult to guarantee the Quality of Service (QoS) for SDN users. To address this issue, this paper proposes a paradigm that uses blockchain to incentive safe load balancing for multiple controllers. We proposed a controller consortium blockchain for secure and efficient load balancing of multi-controllers, which includes a new cryptographic currency balance coin and a novel consensus mechanism Proof-of-Balance (PoB). In addition, we have designed a novel game theory-based incentive mechanism to incentive controllers with tight communication resources to offload tasks to idle controllers. The security analysis and performance simulation results indicate the superiority and effectiveness of the proposed scheme.

Advanced Persistent Threats (APTs) refer to sophisticated, prolonged and multi-step attacks, planned and executed by skilled adversaries targeting government and enterprise networks. Attack graphs' topologies can be leveraged to detect, explain and visualize the progress of such attacks. However, due to the abundance of false-positives, such graphs are usually overwhelmingly large and difficult for an analyst to understand. Graph partitioning refers to the problem of reducing the graph of alerts to a set of smaller incidents that are easier for an analyst to process and better represent the actual attack plan. Existing approaches are oblivious to the security-context of the problem at hand and result in graphs which, while smaller, make little sense from a security perspective. In this paper, we propose an optimization approach allowing us to generate security-aware partitions, utilizing aspects such as the kill chain progression, number of assets involved, as well as the size of the graph. Using real-world datasets, the results show that our approach produces graphs that are better at capturing the underlying attack compared to state-of-the-art approaches and are easier for the analyst to understand.

Network security has become an important issue in our work and life. Hackers' attack mode has been upgraded from normal attack to APT( Advanced Persistent Threat, APT) attack. The key of APT attack chain is the penetration and intrusion of active directory, which can not be completely detected via the traditional IDS and antivirus software. Further more, lack of security protection of existing solutions for domain control aggravates this problem. Although researchers have proposed methods for domain attack detection, many of them have not yet been converted into effective market-oriented products. In this paper, we analyzes the common domain intrusion methods, various domain related attack behavior characteristics were extracted from ATT&CK matrix (Advanced tactics, techniques, and common knowledge) for analysis and simulation test. Based on analyzing the log file generated by the attack, the domain attack detection rules are established and input into the analysis engine. Finally, the available domain intrusion detection system is designed and implemented. Experimental results show that the network attack detection method based on the analysis of domain attack behavior can analyze the log file in real time and effectively detect the malicious intrusion behavior of hackers , which could facilitate managers find and eliminate network security threats immediately.

Security and privacy in computer systems has always been an important aspect of computer engineering and will continue to grow in importance as computer systems become entrusted to handle an ever increasing amount of sensitive information. Classical exploitation techniques such as memory corruption or shell command injection have been well researched and thus there exists known design patterns to avoid and penetration testing tools for testing the robustness of programs against these types of attacks. When it comes to the notion of program security requirements being violated through indirect means referred to as side-channels, testing frameworks of quality comparable to popular memory safety or command injection tools are not available. Recent computer security research has shown that private information may be indirectly leaked through side-channels such as patterns of encrypted network traffic, CPU and motherboard noise, and monitor ambient light. This paper presents the design and evaluation of a side-channel detection and exploitation framework that follows a machine learning based plugin oriented architecture thus allowing side-channel research to be conducted on a wide-variety of side-channel sources.

Accessing the secured data through the network is a major task in emerging technology. Data needs to be protected from the network vulnerabilities, malicious users, hackers, sniffers, intruders. The novel framework has been designed to provide high security in data transaction through computer network. The implant of network amalgamation in the recent trends, make the way in security enhancement in an efficient manner through the machine learning algorithm. In this system the usage of the biometric authenticity plays a vital role for unique approach. The novel mathematical approach is used in machine learning algorithms to solve these problems and provide the security enhancement. The result shows that the novel method has consistent improvement in enhancing the security of data transactions in the emerging technologies.

Network adversaries, such as malicious transit autonomous systems (ASes), have been shown to be capable of partitioning the Bitcoin's peer-to-peer network via routing-level attacks; e.g., a network adversary exploits a BGP vulnerability and performs a prefix hijacking attack (viz. Apostolaki et al. [3]). Due to the nature of BGP operation, such a hijacking is globally observable and thus enables immediate detection of the attack and the identification of the perpetrator. In this paper, we present a stealthier attack, which we call the EREBUS attack, that partitions the Bitcoin network without any routing manipulations, which makes the attack undetectable to control-plane and even to data-plane detectors. The novel aspect of EREBUS is that it makes the adversary AS a natural man-in-the-middle network of all the peer connections of one or more targeted Bitcoin nodes by patiently influencing the targeted nodes' peering decision. We show that affecting the peering decision of a Bitcoin node, which is believed to be infeasible after a series of bug patches against the earlier Eclipse attack [29], is possible for the network adversary that can use abundant network address resources (e.g., spoofing millions of IP addresses in many other ASes) reliably for an extended period of time at a negligible cost. The EREBUS attack is readily available for large ASes, such as Tier-1 and large Tier-2 ASes, against the vast majority of 10K public Bitcoin nodes with only about 520 bit/s of attack traffic rate per targeted Bitcoin node and a modest (e.g., 5-6 weeks) attack execution period. The EREBUS attack can be mounted by nation-state adversaries who would be willing to execute sophisticated attack strategies patiently to compromise cryptocurrencies (e.g., control the consensus, take down a cryptocurrency, censor transactions). As the attack exploits the topological advantage of being a network adversary but not the specific vulnerabilities of Bitcoin core, no quick patches seem to be available. We discuss that some naive solutions (e.g., whitelisting, rate-limiting) are ineffective and third-party proxy solutions may worsen the Bitcoin's centralization problem. We provide some suggested modifications to the Bitcoin core and show that they effectively make the EREBUS attack significantly harder; yet, their non-trivial changes to the Bitcoin's network operation (e.g., peering dynamics, propagation delays) should be examined thoroughly before their wide deployment.

Payment channel networks have been introduced to mitigate the scalability issues inherent to permissionless decentralized cryptocurrencies such as Bitcoin. Launched in 2018, the Lightning Network (LN) has been gaining popularity and consists today of more than 5000 nodes and 35000 payment channels that jointly hold 965 bitcoins (9.2M USD as of June 2020). This adoption has motivated research from both academia and industryPayment channels suffer from security vulnerabilities, such as the wormhole attack [39], anonymity issues [38], and scalability limitations related to the upper bound on the number of concurrent payments per channel [28], which have been pointed out by the scientific community but never quantitatively analyzedIn this work, we first analyze the proneness of the LN to the wormhole attack and attacks against anonymity. We observe that an adversary needs to control only 2% of nodes to learn sensitive payment information (e.g., sender, receiver, and amount) or to carry out the wormhole attack. Second, we study the management of concurrent payments in the LN and quantify its negative effect on scalability. We observe that for micropayments, the forwarding capability of up to 50% of channels is restricted to a value smaller than the channel capacity. This phenomenon hinders scalability and opens the door for denial-of-service attacks: we estimate that a network-wide DoS attack costs within 1.6M USD, while isolating the biggest community costs only 238k USDOur findings should prompt the LN community to consider the issues studied in this work when educating users about path selection algorithms, as well as to adopt multi-hop payment protocols that provide stronger security, privacy and scalability guarantees.

Due to improving computational capacity of supercomputers, transmitting encrypted packets via one single network path is vulnerable to brute-force attacks. The versatile attackers secretly eavesdrop all the packets, classify packets into different streams, performs an exhaustive search for the decryption key, and extract sensitive personal information from the streams. However, new Internet Protocol (IP) brings great opportunities and challenges for preventing eavesdropping attacks. In this paper, we propose a Programming Protocol-independent Packet Processors (P4) based Network Immune Scheme (P4NIS) against the eavesdropping attacks. Specifically, P4NIS is equipped with three lines of defense to improve the network immunity. The first line is promiscuous forwarding by splitting all the traffic packets in different network paths disorderly. Complementally, the second line encrypts transmission port fields of the packets using diverse encryption algorithms. The encryption could distribute traffic packets from one stream into different streams, and disturb eavesdroppers to classify them correctly. Besides, P4NIS inherits the advantages from the existing encryption-based countermeasures which is the third line of defense. Using a paradigm of programmable data planes-P4, we implement P4NIS and evaluate its performances. Experimental results show that P4NIS can increase difficulties of eavesdropping significantly, and increase transmission throughput by 31.7% compared with state-of-the-art mechanisms.

Password Guessing Attacks, for instance, Brute Force and word reference ambushes on online records are directly wide spread. Guarding the ambushes and giving the accommodating login the genuine customers together is a problematic endeavour. The present structures are lacking to give both the security and solace together. Phishing is a digital assault that targets credulous online clients fooling into uncovering delicate data, for example, username, secret key, standardized savings number or charge card number and so forth. Assailants fool the Internet clients by concealing site page as a dependable or real page to recover individual data. Password Guessing Attacks Resistance Protocol (PGARP) limits the full-scale number of logins attempts from darken remote hosts to as low as a single undertaking for each username, genuine customers all around (e.g., when tries are created utilizing known, occasionally used machines) can make a couple failed login tries before being tried with an ATT. A specific most distant point will be made to oblige the number of failed attempts with the ATT in order to keep the attacks. After the failed login attempt with ATT limit accomplished, an admonition will be sent to the customer concerning the failed login tries have accomplished the best measurement. This admonition will caution the customer and the customer will be urged to change the mystery expression and security question.

The Software Defined Network (SDN) provides higher programmable functionality for network configuration and management dynamically. Moreover, SDN introduces a centralized management approach by dividing the network into control and data planes. In this paper, we introduce a deep learning enabled intrusion detection and prevention system (DL-IDPS) to prevent secure shell (SSH) brute-force attacks and distributed denial-of-service (DDoS) attacks in SDN. The packet length in SDN switch has been collected as a sequence for deep learning models to identify anomalous and malicious packets. Four deep learning models, including Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) and Stacked Auto-encoder (SAE), are implemented and compared for the proposed DL-IDPS. The experimental results show that the proposed MLP based DL-IDPS has the highest accuracy which can achieve nearly 99% and 100% accuracy to prevent SSH Brute-force and DDoS attacks, respectively.

Brute-force login attempts are common for every host on the public Internet. While most of them can be discarded as low-threat attacks, targeted attack campaigns often use a dictionary-based brute-force attack to establish a foothold in the network. Therefore, it is important to characterize the attackers' behavior to prioritize defensive measures and react to new threats quickly. In this paper we present a set of metrics that can support threat hunters in characterizing brute-force login attempts. Based on connection metadata, timing information, and the attacker's dictionary these metrics can help to differentiate scans and to find common behavior across distinct IP addresses. We evaluated our novel metrics on a real-world data set of malicious login attempts collected by our honeypot Honeygrove. We highlight interesting metrics, show how clustering can be leveraged to reveal common behavior across IP addresses, and describe how selected metrics help to assess the threat level of attackers. Amongst others, we for example found strong indicators for collusion between ten otherwise unrelated IP addresses confirming that a clustering of the right metrics can help to reveal coordinated attacks.

Network traffic anomaly detection is of critical importance in cybersecurity due to the massive and rapid growth of sophisticated computer network attacks. Indeed, the more new Internet-related technologies are created, the more elaborate the attacks become. Among all the contemporary high-level attacks, dictionary-based brute-force attacks (BFA) present one of the most unsurmountable challenges. We need to develop effective methods to detect and mitigate such brute-force attacks in realtime. In this paper, we investigate SSH and FTP brute-force attack detection by using the Long Short-Term Memory (LSTM) deep learning approach. Additionally, we made use of machine learning (ML) classifiers: J48, naive Bayes (NB), decision table (DT), random forest (RF) and k-nearest-neighbor (k-NN), for additional detection purposes. We used the well-known labelled dataset CICIDS2017. We evaluated the effectiveness of the LSTM and ML algorithms, and compared their performance. Our results show that the LSTM model outperforms the ML algorithms, with an accuracy of 99.88%.

P2P botnet has become one of the most serious threats to today's network security. It can be used to launch kinds of malicious activities, ranging from spamming to distributed denial of service attack. However, the detection of P2P botnet is always challenging because of its decentralized architecture. In this paper, we propose a two-stage P2P botnet detection method which only relies on several traffic statistical features. This method first detects P2P hosts based on three statistical features, and then distinguishes P2P bots from benign P2P hosts by means of another two statistical features. Experimental evaluations on real-world traffic datasets shows that our method is able to detect hidden P2P bots with a detection accuracy of 99.7% and a false positive rate of only 0.3% within 5 minutes.

Botnet is one of the threats to internet network security-Botmaster in carrying out attacks on the network by relying on communication on network traffic. Internet of Things (IoT) network infrastructure consists of devices that are inexpensive, low-power, always-on, always connected to the network, and are inconspicuous and have ubiquity and inconspicuousness characteristics so that these characteristics make IoT devices an attractive target for botnet malware attacks. In identifying whether packet traffic is a malware attack or not, one can use machine learning classification methods. By using Weka and Scikit-learn analysis tools machine learning, this paper implements four machine learning algorithms, i.e.: AdaBoost, Decision Tree, Random Forest, and Naïve Bayes. Then experiments are conducted to measure the performance of the four algorithms in terms of accuracy, execution time, and false positive rate (FPR). Experiment results show that the Weka tool provides more accurate and efficient classification methods. However, in false positive rate, the use of Scikit-learn provides better results.

Internet of Things (IoT) is a revolutionary expandable network which has brought many advantages, improving the Quality of Life (QoL) of individuals. However, IoT carries dangers, due to the fact that hackers have the ability to find security gaps in users' IoT devices, which are not still secure enough and hence, intrude into them for malicious activities. As a result, they can control many connected devices in an IoT network, turning IoT into Botnet of Things (BoT). In a botnet, hackers can launch several types of attacks, such as the well known attacks of Distributed Denial of Service (DDoS) and Man in the Middle (MitM), and/or spread various types of malicious software (malware) to the compromised devices of the IoT network. In this paper, we propose a novel hybrid Artificial Intelligence (AI)-powered honeynet for enhanced IoT botnet detection rate with the use of Cloud Computing (CC). This upcoming security mechanism makes use of Machine Learning (ML) techniques like the Logistic Regression (LR) in order to predict potential botnet existence. It can also be adopted by other conventional security architectures in order to intercept hackers the creation of large botnets for malicious actions.

Today our world benefits from Internet of Things (IoT) technology; however, new security problems arise when these IoT devices are introduced into our homes. Because many of these IoT devices have access to the Internet and they have little to no security, they make our smart homes highly vulnerable to compromise. Some of the threats include IoT botnets and generic confidentiality, integrity, and availability (CIA) attacks. Our research explores botnet detection by experimenting with supervised machine learning and deep-learning classifiers. Further, our approach assesses classifier performance on unbalanced datasets that contain benign data, mixed in with small amounts of malicious data. We demonstrate that the classifiers can separate malicious activity from benign activity within a small IoT network dataset. The classifiers can also separate malicious activity from benign activity in increasingly larger datasets. Our experiments have demonstrated incremental improvement in results for (1) accuracy, (2) probability of detection, and (3) probability of false alarm. The best performance results include 99.9% accuracy, 99.8% probability of detection, and 0% probability of false alarm. This paper also demonstrates how the performance of these classifiers increases, as IoT training datasets become larger and larger.

Reliability analysis of concurrent data based on Botnet modeling is conducted in this paper. At present, the detection methods for botnets are mainly focused on two aspects. The first type requires the monitoring of high-privilege systems, which will bring certain security risks to the terminal. The second type is to identify botnets by identifying spam or spam, which is not targeted. By introducing multi-dimensional permutation entropy, the impact of permutation entropy on the permutation entropy is calculated based on the data communicated between zombies, describing the complexity of the network traffic time series, and the clustering variance method can effectively solve the difficulty of the detection. This paper is organized based on the data complex structure analysis. The experimental results show acceptable performance.

Through analysis of sessions in which files were created and downloaded on three Cowrie SSH/Telnet honeypots, we find that IoT botnets are by far the most common source of malware on connected systems with weak credentials. We detail our honeypot configuration and describe a simple method for listing near-identical malicious login sessions using edit distance. A large number of IoT botnets attack our honeypots, but the malicious sessions which download botnet software to the honeypot are almost all nearly identical to one of two common attack patterns. It is apparent that the Mirai worm is still the dominant botnet software, but has been expanded and modified by other hackers. We also find that the same loader devices deploy several different botnet malware strains to the honeypot over the course of a 40 day period, suggesting multiple botnet deployments from the same source. We conclude that Mirai continues to be adapted but can be effectively tracked using medium interaction honeypots such as Cowrie.

Among the different types of malware, botnets are rising as the most genuine risk against cybersecurity as they give a stage to criminal operations (e.g., Distributed Denial of Service (DDOS) attacks, malware dispersal, phishing, and click fraud and identity theft). Existing botnet detection techniques work only on specific botnet Command and Control (C&C) protocols and lack in providing early-stage botnet detection. In this paper, we propose an approach for early-stage botnet detection. The proposed approach first selects the optimal features using feature selection techniques. Next, it feeds these features to machine learning classifiers to evaluate the performance of the botnet detection. Experiments reveals that the proposed approach efficiently classifies normal and malicious traffic at an early stage. The proposed approach achieves the accuracy of 99%, True Positive Rate (TPR) of 0.99 %, and False Positive Rate (FPR) of 0.007 % and provide an efficient detection rate in comparison with the existing approach.

This paper proposes a deep learning-based white-hat worm launcher in Botnet Defense System (BDS). BDS uses white-hat botnets to defend an IoT system against malicious botnets. White-hat worm launcher literally launches white-hat worms to create white-hat botnets according to the strategy decided by BDS. The proposed launcher learns with deep learning where is the white-hat worms' right place to successfully drive out malicious botnets. Given a system situation invaded by malicious botnets, it predicts a worms' placement by the learning result and launches them. We confirmed the effect of the proposed launcher through simulating evaluation.

Botnets are one of the major threats on the Internet. They are used for malicious activities to compromise the basic network security goals, namely Confidentiality, Integrity, and Availability. For reliable botnet detection and defense, deep learning-based approaches were recently proposed. In this paper, four different deep learning models, namely Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), hybrid CNN-LSTM, and Multi-layer Perception (MLP) are applied for botnet detection and simulation studies are carried out using the CTU-13 botnet traffic dataset. We use several performance metrics such as accuracy, sensitivity, specificity, precision, and F1 score to evaluate the performance of each model on classifying both known and unknown (zero-day) botnet traffic patterns. The results show that our deep learning models can accurately and reliably detect both known and unknown botnet traffic, and show better performance than other deep learning models.

The increase of cyber attacks in both the numbers and varieties in recent years demands to build a more sophisticated network intrusion detection system (NIDS). These NIDS perform better when they can monitor all the traffic traversing through the network like when being deployed on a Software-Defined Network (SDN). Because of the inability to detect zero-day attacks, signature-based NIDS which were traditionally used for detecting malicious traffic are beginning to get replaced by anomaly-based NIDS built on neural networks. However, recently it has been shown that such NIDS have their own drawback namely being vulnerable to the adversarial example attack. Moreover, they were mostly evaluated on the old datasets which don't represent the variety of attacks network systems might face these days. In this paper, we present Reconstruction from Partial Observation (RePO) as a new mechanism to build an NIDS with the help of denoising autoencoders capable of detecting different types of network attacks in a low false alert setting with an enhanced robustness against adversarial example attack. Our evaluation conducted on a dataset with a variety of network attacks shows denoising autoencoders can improve detection of malicious traffic by up to 29% in a normal setting and by up to 45% in an adversarial setting compared to other recently proposed anomaly detectors.

The legacy security defense mechanisms cannot resist where emerging sophisticated threats such as zero-day and malware campaigns have profoundly changed the dimensions of cyber-attacks. Recent studies indicate that cyber threat intelligence plays a crucial role in implementing proactive defense operations. It provides a knowledge-sharing platform that not only increases security awareness and readiness but also enables the collaborative defense to diminish the effectiveness of potential attacks. In this paper, we propose a secure distributed model to facilitate cyber threat intelligence sharing among diverse participants. The proposed model uses blockchain technology to assure tamper-proof record-keeping and smart contracts to guarantee immutable logic. We use an open-source permissioned blockchain platform, Hyperledger Fabric, to implement the blockchain application. We also utilize the flexibility and management capabilities of Software-Defined Networking to be integrated with the proposed sharing platform to enhance defense perspectives against threats in the system. In the end, collaborative DDoS attack mitigation is taken as a case study to demonstrate our approach.