Search

Nowadays, online services, like e-commerce or streaming services, provide a personalized user experience through recommender systems. Recommender systems are built upon a vast amount of data about users/items acquired by the services. Such knowledge represents an invaluable resource. However, commonly, part of this knowledge is public and can be easily accessed via the Internet. Unfortunately, that same knowledge can be leveraged by competitors or malicious users. The literature offers a large number of works concerning attacks on recommender systems, but most of them assume that the attacker can easily access the full rating matrix. In practice, this is never the case. The only way to access the rating matrix is by gathering the ratings (e.g., reviews) by crawling the service's website. Crawling a website has a cost in terms of time and resources. What is more, the targeted website can employ defensive measures to detect automatic scraping. In this paper, we assess the impact of a series of attacks on recommender systems. Our analysis aims to set up the most realistic scenarios considering both the possibilities and the potential attacker's limitations. In particular, we assess the impact of different crawling approaches when attacking a recommendation service. From the collected information, we mount various profile injection attacks. We measure the value of the collected knowledge through the identification of the most similar user/item. Our empirical results show that while crawling can indeed bring knowledge to the attacker (up to 65% of neighborhood reconstruction on a mid-size dataset and up to 90% on a small-size dataset), this will not be enough to mount a successful shilling attack in practice. [ABSTRACT FROM AUTHOR]

A distributed denial of service (DDoS) attack on any of the major components (e.g., controller, switches, and southbound channel) of software defined networking (SDN) architecture is a critical security threat. For example, the breakdown of controller could disrupt the data communication in the whole SDN network. A possible way to perform DoS is to generate a large number of new, but short length traffic flows. These flows will trigger malicious flooding requests to overload the controller and causes overflow in flow tables at SDN switches. In this paper, we propose two lightweight and practically feasible countermeasures against two different types of DDoS attacks called Route Spoofing and Resource Exhaustion in SDN networks. For Route Spoofing attack, we introduce a technique called "selective blocking", which stops an adversary node from maliciously using other users active communication routes. To countermeasure Resource Exhaustion attack, we propose a solution called "periodic monitoring", which detects adversary nodes based on the traffic analysis statistics that are gathered within a time window. We implement and perform result analysis of the attacks and their proposed countermeasures. When using our proposed countermeasures in the target SDN scenarios, the simulation results indicate an adequate reduction in bandwidth consumption and processing delay of new request, and it also depicts substantial gain in packet delivery rate. Additionally, we present the receiver operating characteristic curve, which shows the sensitivity and specificity of our countermeasures along with their detection accuracy. [ABSTRACT FROM AUTHOR]

Enforcing security often requires the two legitimate parties of a communication to determine if they share a secret, without disclosing information (e.g., the shared secret itself, or just the existence of such a secret) to third parties. In this paper, we propose CASSANDRA, a toolbox composed of three probabilistic protocols that allows two parties, each one having a subset of elements drawn by a pre-determined set, to compute information about the intersection of such two sets. In particular, C-void decides whether the two sets are disjoint; C-size allows to compute how many elements the intersection is composed of; and, C-set returns the identity of the elements of the intersection (if any). These protocols differ, other than in functionality, also in the degree of assurance they can provide and the degree of interactions required by the two parties. The communication cost also differs, but in any case, it is below the cost of competing solution representing the state of the art. These protocols also share some common features: that is, they are completely tunable and specifically suited for devices having constraints on energy, communication, storage, and bandwidth. Examples of these devices are portable devices (e.g., phones) handling satellite communications, or nodes of wireless sensor networks. Thorough analysis and extensive simulations support our findings. [ABSTRACT FROM AUTHOR]

Summary: In software‐defined networking (SDN), TCP SYN flooding attack is considered as one of the most effective attacks to perform control plane and target server saturation. In this attack, an attacker generates a large number of malicious SYN requests, and because of the absence of the forwarding rules, the data plane switches have to forward these SYN messages to the controller. This excessive forwarding causes congestion over the communication channel between a data plane and control plane, and it also exhausts computational resources at both the planes. In this paper, we propose a novel countermeasure called SYN‐Guard to detect and prevent SYN flooding in SDN networks. We fully implement SYN‐Guard on the SDN controller to validate the incoming TCP connection requests. The controller installs forwarding rules for the SYN requests that successfully clear the validation test of SYN‐Guard. The host of the fake SYN request is detected, and SYN‐Guard prevents it from sending any further SYN requests to the data plane switch. The performance evaluation done using the simulation results shows that SYN‐Guard exhibits low side effect for genuine TCP requests, and when compared with standard SDN and state‐of‐art proposals, it reduces the average response time up to 21% during an ongoing SYN flooding attack. [ABSTRACT FROM AUTHOR]

Android is currently the most widespread operating system (OS) worldwide, but also the most prone to attacks. Despite the challenges faced by Industry and Academia to improve the Android OS security, it still has several vulnerabilities. Among those, the severity of the Next-Intent Vulnerability (NIV) can be immediately grasped. Android apps are made of components, which by default are private and cannot be targeted by other apps on the same phone. However, NIV allows any app to access the private components of a different app, eventually generating a crash or stealing sensitive data. NIV occurs when there is a chain of calls among different components based on the Intent messaging model and there is no control over the reliability of the first component triggering the call. NIV was first detected in 2013, but it is still an open issue. In this paper, we present Next-Intent Vulnerability Detector ( N I V D ), a novel approach to detect NIV in Android apps by relying on type systems. N I V D applies the inference rules of its type system to the app execution paths containing a sequence of calls to three NIV-related Android APIs. Compared to the state-of-the-art, N I V D is faster and more efficient, without losing precision in detecting NIV. Finally, through N I V D Google Photos was found to be vulnerable, and we disclosed the finding on the Google official bug report website (issue number 124342801). [ABSTRACT FROM AUTHOR]

In this paper, we show a deep learning model for bot detection, named as DeBot, for industrial network traffic. DeBot uses a novel Cascade Forward Back Propagation Neural Network (CFBPNN) model with a subset of features using the Correlation-based Feature Selection (CFS) technique. A time series-based Nonlinear Auto-regressive Network with eXogenous inputs (NARX) technique analyzes the factors having a higher impact on the target variable and predict the behavioral pattern. To the best of our knowledge, we pioneer the use of optimal feature selection and integration with the cascading model of deep learning in bot detection of IIoTs. We conduct a thorough set of experiments on five popular bot datasets: NF-UNSW-NB15, NF-ToN-IoT, NF-BoT-IoT, NF-CSE-CIC-IDS2018, and ToN-IoT-Windows. We compare CFBPNN with other existing neural network models. We observe that CFBPNN in DeBot shows 100% accuracy in all the datasets with subset evaluation and obtains optimum F1-score and zero precision. [Display omitted] • We address the multi-layered network model with IIoT features to enable bot attack detection. • Our DeBot model evaluates the subset with optimal features with high impact on attack class and predict the risk factor. • We pioneer the implementation of time series-based Nonlinear Auto-Regressive eXogenous (NARX) model to predict the risk in the specified time gap. • We introduce a new cascading model with recurrent feature to preserve the relationship between input and output variables. [ABSTRACT FROM AUTHOR]