Search for: All records

As the web keeps increasing in size, the number of vulnerable and poorly-managed websites increases commensurately. Attackers rely on armies of malicious bots to discover these vulnerable websites, compromising their servers, and exfiltrating sensitive user data. It is therefore crucial for the security of the web to understand the population and behavior of malicious bots. In this paper, we report on the design, implementation, and results of Aristaeus, a system for deploying large numbers of honeysites, i.e., websites that exist for the sole purpose of attracting and recording bot traffic. Through a seven-month-long experiment with 100 dedicated honeysites, Aristaeus recorded 26.4 million requests sent by more than 287K unique IP addresses, with 76K of them belonging to clearly malicious bots. By analyzing the type of requests and payloads that these bots send, we discover that the average honeysite received more than 37K requests each month, with more than 50% of these requests attempting to brute-force credentials, fingerprint the deployed web applications, and exploit large numbers of different vulnerabilities. By comparing the declared identity of these bots with their TLS handshakes and HTTP headers, we uncover that more than 86.2% of bots claiming to be Mozilla Firefox and Google Chrome are lying about their identity and are instead built on HTTP libraries and command-line tools. 

Serverless Computing has quickly emerged as a dominant cloud computing paradigm, allowing developers to rapidly prototype event-driven applications using a composition of small functions that each perform a single logical task. However, many such application workflows are based in part on publicly-available functions developed by third-parties, creating the potential for functions to behave in unexpected, or even malicious, ways. At present, developers are not in total control of where and how their data is flowing, creating significant security and privacy risks in growth markets that have embraced serverless (e.g., IoT). As a practical means of addressing this problem, we present Valve, a serverless platform that enables developers to exert complete fine-grained control of information flows in their applications. Valve enables workflow developers to reason about function behaviors, and specify restrictions, through auditing of network-layer information flows. By proxying network requests and propagating taint labels across network flows, Valve is able to restrict function behavior without code modification. We demonstrate that Valve is able defend against known serverless attack behaviors including container reuse-based persistence and data exfiltration over cloud platform APIs with less than 2.8% runtime overhead, 6.25% deployment overhead and 2.35% teardown overhead. 

Emerging smart home platforms, which interface with a variety of physical devices and support third-party application development, currently use permission models inspired by smartphone operating systems—the permission to access operations are separated by the device which performs them instead of their functionality. Unfortunately, this leads to two issues: (1) apps that do not require access to all of the granted device operations have overprivileged access to them, (2) apps might pose a higher risk to users than needed because physical device operations are fundamentally risk-asymmetric — “door.unlock” provides access to burglars, and “door.lock” can potentially lead to getting locked out. Overprivileged apps with access to mixed-risk operations only increase the potential for damage. We present Tyche, a secure development methodology that leverages the risk-asymmetry in physical device operations to limit the risk that apps pose to smart home users, without increasing the user’s decision overhead. Tyche introduces the notion of risk-based permissions for IoT systems. When using risk-based permissions, device operations are grouped into units of similar risk, and users grant apps access to devices at that risk-based granularity. Starting from a set of permissions derived from the popular Samsung SmartThings platform, we conduct a user study involving domain-experts and Mechanical Turk users to compute a relative ranking of risks associated with device operations. We find that user assessment of risk closely matches that of domain experts. Using this insight, we define risk-based groupings of device operations, and apply it to existing SmartThings apps. We show that existing apps can reduce access to high-risk operations by 60% while remaining operable. 

Trigger-Action platforms are web-based systems that enable users to create automation rules by stitching together online services representing digital and physical resources using OAuth tokens. Unfortunately, these platforms introduce a longrange large-scale security risk: If they are compromised, an attacker can misuse the OAuth tokens belonging to a large number of users to arbitrarily manipulate their devices and data. We introduce Decentralized Action Integrity, a security principle that prevents an untrusted trigger-action platform from misusing compromised OAuth tokens in ways that are inconsistent with any given user’s set of trigger-action rules. We present the design and evaluation of Decentralized Trigger-Action Platform (DTAP), a trigger-action platform that implements this principle by overcoming practical challenges. DTAP splits currently monolithic platform designs into an untrusted cloud service, and a set of user clients (each user only trusts their client). Our design introduces the concept of Transfer Tokens (XTokens) to practically use finegrained rule-specific tokens without increasing the number of OAuth permission prompts compared to current platforms. Our evaluation indicates that DTAP poses negligible overhead: it adds less than 15ms of latency to rule execution time, and reduces throughput by 2.5%. 

Deep neural networks (DNNs) are vulnerable to adversarial examples—maliciously crafted inputs that cause DNNs to make incorrect predictions. Recent work has shown that these attacks generalize to the physical domain, to create perturbations on physical objects that fool image classifiers under a variety of real-world conditions. Such attacks pose a risk to deep learning models used in safety-critical cyber-physical systems. In this work, we extend physical attacks to more challenging object detection models, a broader class of deep learning algorithms widely used to detect and label multiple objects within a scene. Improving upon a previous physical attack on image classifiers, we create perturbed physical objects that are either ignored or mislabeled by object detection models. We implement a Disappearance Attack, in which we cause a Stop sign to “disappear” according to the detector—either by covering the sign with an adversarial Stop sign poster, or by adding adversarial stickers onto the sign. In a video recorded in a controlled lab environment, the state-of-the-art YOLO v2 detector failed to recognize these adversarial Stop signs in over 85% of the video frames. In an outdoor experiment, YOLO was fooled by the poster and sticker attacks in 72.5% and 63.5% of the video frames respectively. We also use Faster R-CNN, a different object detection model, to demonstrate the transferability of our adversarial perturbations. The created poster perturbation is able to fool Faster R-CNN in 85.9% of the video frames in a controlled lab environment, and 40.2% of the video frames in an outdoor environment. Finally, we present preliminary results with a new Creation Attack, wherein innocuous physical stickers fool a model into detecting nonexistent objects. 

Recent studies show that the state-of-the-art deep neural networks (DNNs) are vulnerable to adversarial examples, resulting from small-magnitude perturbations added to the input. Given that that emerging physical systems are using DNNs in safety-critical situations, adversarial examples could mislead these systems and cause dangerous situations. Therefore, understanding adversarial examples in the physical world is an important step towards developing resilient learning algorithms. We propose a general attack algorithm, Robust Physical Perturbations (RP2), to generate robust visual adversarial perturbations under different physical conditions. Using the real-world case of road sign classification, we show that adversarial examples generated using RP2 achieve high targeted misclassification rates against standard-architecture road sign classifiers in the physical world under various environmental conditions, including viewpoints. Due to the current lack of a standardized testing method, we propose a two-stage evaluation methodology for robust physical adversarial examples consisting of lab and field tests. Using this methodology, we evaluate the efficacy of physical adversarial manipulations on real objects. With a perturbation in the form of only black and white stickers, we attack a real stop sign, causing targeted misclassification in 100% of the images obtained in lab settings, and in 84.8% of the captured video frames obtained on a moving vehicle (field test) for the target classifier. 

Recent studies show that the state-of-the-art deep neural networks (DNNs) are vulnerable to adversarial examples, resulting from small-magnitude perturbations added to the input. Given that that emerging physical systems are using DNNs in safety-critical situations, adversarial examples could mislead these systems and cause dangerous situations. Therefore, understanding adversarial examples in the physical world is an important step towards developing resilient learning algorithms. We propose a general attack algorithm, Robust Physical Perturbations (RP 2 ), to generate robust visual adversarial perturbations under different physical conditions. Using the real-world case of road sign classification, we show that adversarial examples generated using RP 2 achieve high targeted misclassification rates against standard-architecture road sign classifiers in the physical world under various environmental conditions, including viewpoints. Due to the current lack of a standardized testing method, we propose a two-stage evaluation methodology for robust physical adversarial examples consisting of lab and field tests. Using this methodology, we evaluate the efficacy of physical adversarial manipulations on real objects. With a perturbation in the form of only black and white stickers, we attack a real stop sign, causing targeted misclassification in 100% of the images obtained in lab settings, and in 84.8% of the captured video frames obtained on a moving vehicle (field test) for the target classifier. 

Many of the everyday decisions a user makes rely on the suggestions of online recommendation systems. These systems amass implicit (e.g., location, purchase history, browsing history) and explicit (e.g., reviews, ratings) feedback from multiple users, produce a general consensus, and provide suggestions based on that consensus. However, due to privacy concerns, users are uncomfortable with implicit data collection, thus requiring recommendation systems to be overly dependent on explicit feedback. Unfortunately, users do not frequently provide explicit feedback. This hampers the ability of recommendation systems to provide high-quality suggestions. We introduce Heimdall, the first privacy-respecting implicit preference collection framework that enables recommendation systems to extract user preferences from their activities in a privacy respect- ing manner. The key insight is to enable recommendation systems to run a collector on a user’s device and precisely control the information a collector transmits to the recommendation system back- end. Heimdall introduces immutable blobs as a mechanism to guarantee this property. We implemented Heimdall on the Android plat- form and wrote three example collectors to enhance recommendation systems with implicit feedback. Our performance results suggest that the overhead of immutable blobs is minimal, and a user study of 166 participants indicates that privacy concerns are significantly less when collectors record only specific information—a property that Heimdall enables.