Helmholtz Center for Information Security
CISPA – Helmholtz-Zentrum für InformationssicherheitNot a member yet
3406 research outputs found
Sort by
Jack-in-the-box: An Empirical Study of JavaScript Bundling on the Web and its Security Implications
Full text linkIn recent years, we have seen an increased interest in studying the software supply chain of user-facing applications to uncover problematic third-party dependencies. Prior work shows that web applications often rely on outdated or vulnerable third-party code. Moreover, real-world supply chain attacks show that dependencies can also be used to deliver malicious code, e.g., for carrying cryptomining operations. Nonetheless, existing measurement studies in this domain neglect an important software engineering practice: developers often merge together third-party code into a single file called bundle, which they then deliver from their own servers, making it appear as first-party code. Bundlers like Webpack or Rollup are popular open-source projects with tens of thousand of GitHub stars, suggesting that this technology is widely-used by developers. Ignoring bundling may result in underestimating the complexity of modern software supply chains. In this work, we aim to address this methodological shortcomings of prior work. To this end, we propose a novel methodology for automatically detecting bundles, and partially reverse engineer them. Using this methodology, we conduct the first large-scale empirical study of bundled code on the web and examine its security implications. We provide evidence about the high prevalence of bundles, which are contained in 40% of all websites and the average website includes more than one bundle. Following our methodology, we reidentify 1051 vulnerabilities originating from 33 vulnerable npm packages, included in bundled code. Among the vulnerabilities, we find 17 critical and 59 high severity ones, which might enable malicious actors to execute attacks such as arbitrary code execution. Analyzing the low-rated libraries included in bundles, we discover 10 security placeholder packages, which suggest that supply-chain attacks against bundles are not only possible, but they are already happening
VENOMAVE: Targeted Poisoning Against Speech Recognition
No full textDespite remarkable improvements, automatic speech recognition is susceptible to adversarial perturbations. Compared to standard machine learning architectures, these attacks against speech recognition are significantly more challenging, especially since the inputs to a speech recognition system are time series that contain both acoustic and linguistic properties of speech. Extracting all recognition-relevant information requires more complex pipelines and an ensemble of specialized components. Consequently, an attacker needs to consider the entire pipeline.
In this paper, we present VENOMAVE, the first training time poisoning attack against speech recognition. Similar to the pre-dominantly studied evasion attacks, we pursue the same goal: leading the system to an incorrect, and attacker-chosen transcription of a target audio waveform. In contrast to evasion attacks, however, we assume that the attacker can only manipulate a small part of the training data without altering the target audio waveform at run time. We evaluate our attack on two datasets: TIDIGITS and Speech Commands. When poisoning less than 0.17 % of the dataset, VENOMAVE achieves attack success rates of over 80.0 %, without access to the victim’s network architecture or hyperparameters. In a more realistic scenario, when the target audio waveform is played over the air in different rooms, VENOMAVE maintains a success rate of up to 73.3%. Finally, VENOMAVE achieves an attack transferability rate of
36.4 % between two different model architecture
Taming Large Bounds in Synthesis from Bounded-Liveness Specifications
Full text linkAutomatic synthesis from temporal logic specifications is an attractive alternative to manual system design, due to its ability to generate correct-by-construction implementations from high-level specifications. Due to the high complexity of the synthesis problem, significant research efforts have been directed at developing practically efficient approaches for restricted specification language fragments.
In this paper, we focus on the Safety LTL fragment of Linear Temporal Logic (LTL) syntactically extended with bounded temporal operators. We propose a new synthesis approach with the primary motivation to solve efficiently the synthesis problem for specifications with bounded temporal operators, in particular those with large bounds. The experimental evaluation of our method shows that for this type of specifications, it outperforms state-of-art synthesis tools, demonstrating that it is a promising approach to efficiently treating quantitative timing constraints in safety specifications
Bilingual Problems: Studying the Security Risks Incurred by Native Extensions in Scripting Languages
Full text linkScripting languages are continuously gaining popularity due to their ease of use and the flourishing software ecosystems that surround them. These languages offer crash and memory safety by design, thus, developers do not need to understand and prevent low-level security issues like the ones plaguing the C code. However, scripting languages often allow native extensions, which are a way for custom C/C++ code to be invoked directly from the high-level language. While this feature promises several benefits such as increased performance or the reuse of legacy code, it can also break the language's guarantees, e.g., crash-safety.
In this work, we first provide a comparative analysis of the security risks of native extension APIs in three popular scripting languages. Additionally, we discuss a novel methodology for studying the misuse of the native extension API. We then perform an in-depth study of npm, an ecosystem which is most exposed to threats introduced by native extensions. We show that vulnerabilities in extensions can be exploited in their embedding library by producing reads of uninitialized memory, hard crashes or memory leaks in 33 npm packages, simply by invoking their API with well-crafted inputs. Moreover, we identify six open-source web applications in which such exploits can be deployed remotely by a weak adversary. Finally, we were assigned seven security advisories for the work presented in this paper, most labeled as high severity
Drone Security and the Mysterious Case of DJI's DroneID
No full textConsumer drones enable high-class aerial video photography, promise to reform the logistics industry, and are already used for humanitarian rescue operations and during armed conflicts. Contrasting their widespread adoption and high popularity, the low entry barrier for air mobility - a traditionally heavily regulated sector - poses many risks to safety, security, and privacy. Malicious parties could, for example, (mis-)use drones for surveillance, transportation of illegal goods, or cause economic damage by intruding the closed airspace over airports. To prevent harm, drone manufacturers employ several countermeasures to enforce safe and secure use of drones, e.g., they impose software limits regarding speed and altitude, or use geofencing to implement no-fly zones around airports or prisons. Complementing traditional countermeasures, drones from the market leader DJI implement a protocol called DroneID, which is designed to transmit the position of both the drone and its operator to authorized entities such as law enforcement or operators of critical infrastructures.
In this paper, we analyze security and privacy claims for drones, focusing on the leading manufacturer DJI with a market share of 94%. We first systemize the drone attack surface and investigate an attacker capable of eavesdropping on the drone's over-the-air data traffic. Based on reverse engineering of DJI firmware, we design and implement a decoder for DJI's proprietary tracking protocol DroneID using only cheap COTS hardware. We show that the transmitted data is not encrypted, but accessible to anyone, compromising the drone operator's privacy. Second, we conduct a comprehensive analysis of drone security: Using a combination of reverse engineering, a novel fuzzing approach tailored to DJI's communication protocol, and hardware analysis, we uncover several critical flaws in drone firmware that allow attackers to gain elevated privileges on two different DJI drones and their remote control. Such root access paves the way to disable or bypass countermeasures and abuse drones. These vulnerabilities have the potential to be triggered remotely, causing the drone to crash mid-flight
The Leaky Web: Automated Discovery of Cross-Site Information Leaks in Browsers and the Web
Full text linkWhen browsing the web, none of us want sites to infer which other sites we may have visited before or are logged in to. However, attacker-controlled sites may infer this state through browser side-channels dubbed Cross-Site Leaks (XS-Leaks). Although these issues have been known since the 2000s, prior reports mostly found individual instances of issues rather than systematically studying the problem space. Further, actual impact in the wild often remained opaque. To address these open problems, we develop the first automated framework to systematically discover observation channels in browsers. In doing so, we detect and characterize 280 observation channels that leak information cross-site in the engines of Chromium, Firefox, and Safari, which include many variations of supposedly fixed leaks. Atop this framework, we create an automatic pipeline to find XS-Leaks in real-world websites. With this pipeline, we conduct the largest to-date study on XS-Leak prevalence in the wild by performing visit inference and a newly proposed variant cookie acceptance inference attack on the Tranco Top10K. In addition, we test 100 websites for the classic XS-Leak attack vector of login detection. Our results show that XS-Leaks pose a significant threat to the web ecosystem as at least 15%, 34%, and 77% of all tested sites are vulnerable to the three attacks. Also, we present substantial implementation differences between the browsers resulting in differing attack surfaces that matter in the wild. To ensure browser vendors and web developers alike can check their applications for XS-Leaks, we open-source our framework and include an extensive discussion on countermeasures to get rid of XS-Leaks in the near future and ensure new features in browsers do not introduce new XS-Leaks
Special Properties of Gradient Descent with Large Learning Rates
Full text linkWhen training neural networks, it has been widely observed that a large step size is essential in stochastic gradient descent (SGD) for obtaining superior models. However, the effect of large step sizes on the success of SGD is not well understood theoretically. Several previous works have attributed this success to the stochastic noise present in SGD. However, we show through a novel set of experiments that the stochastic noise is not sufficient to explain good non-convex training, and that instead the effect of a large learning rate itself is essential for obtaining best performance.We demonstrate the same effects also in the noise-less case, i.e. for full-batch GD. We formally prove that GD with large step size ---on certain non-convex function classes --- follows a different trajectory than GD with a small step size, which can lead to convergence to a global minimum instead of a local one. Our settings provide a framework for future analysis which allows comparing algorithms based on behaviors that can not be observed in the traditional settings
Why Random Pruning Is All We Need to Start Sparse
Full text linkRandom masks define surprisingly effective sparse neural network models, as has been shown empirically. The resulting sparse networks can often compete with dense architectures and state-of-the-art lottery ticket pruning algorithms, even
though they do not rely on computationally expensive prune-train iterations and can be drawn initially without significant computational over-head. We offer a theoretical explanation of how random masks can approximate arbitrary target networks if they are wider by a logarithmic factor
in the inverse sparsity 1/ log(1/sparsity). This overparameterization factor is necessary at least for 3-layer random networks, which elucidates the observed degrading performance of random networks at higher sparsity. At moderate to high sparsity levels, however, our results imply that sparser networks are contained within random source networks so that any dense-to-sparse training scheme can be turned into a computationally more efficient sparse to sparse one by constraining the search to a fixed random mask. We demonstrate the feasibility of this approach
in experiments for different pruning methods and propose particularly effective choices of initial layer-wise sparsity ratios of the random source network. As a special case, we show theoretically and experimentally that random source networks also contain strong lottery tickets. Our code is available at https://github.com/RelationalML/sparse_to_sparse
Perceptions of Distributed Ledger Technology Key Management - An Interview Study with Finance Professionals
Full text linkKey management is an integral part of using distributed ledger technology (DLT). Previous work has primarily focused on key management for single-user scenarios on Bitcoin. Over the last decade, DLT has evolved to commercial and financial sectors; for example, a new German law allows the trading of a variety of financial securities via DLT. Instead of a single-user paradigm, financial institutions follow a multi-user paradigm. Combining multi-user key management with single-user key management solutions leads to unique challenges with usability and security.
We extend current research through a two-stage qualitative interview study with 13 finance professionals. We investigate how the technical reality contrasts with perceptions of key management practices in corporate financial organizations. Our interdisciplinary study shows, among other things, that DLT does not meet real-world requirements in this particular domain. Moreover, it introduces additional challenges in terms of authentication and auditing.
Our findings suggest that corporate financial institutions strongly support the adoption of blockchain solutions. However, to comply with regulatory and operational requirements, they face additional usability and security challenges, e.g., authentication and access control. Better mechanisms or novel design approaches are required to cover professional environments. This includes how multiple users can access the same assets and approve joint transactions
Practical Timing Side-Channel Attacks on Memory Compression
Full text linkCompression algorithms have side channels due to their data-dependent operations. So far only the compression-ratio side channel was exploited, e.g., the compressed data size. In this paper, we present Decomp+Time, the first memory compression attack exploiting a timing side channel in compression algorithms. While Decomp+Time affects a much broader set of applications than prior work, a key challenge is precisely crafting attacker-controlled compression payloads to enable the attack with sufficient resolution. We develop an evolutionary fuzzer, Comprezzor, to find effective Decomp+Time payloads that optimize latency differences and find payloads that are so effective that decompression timing can even be exploited in remote Decomp+Time attacks across the Internet. Decomp+Time has a capacity of 9.73 kB/s locally, and 10.72 bit/min across the internet (14 hops, > 700 miles). Using Comprezzor, we develop attacks that leak data byte-by-byte in four different case studies: First, we leak 1.50 bit/min from Memcached on a remote server running a PHP application. Second, we leak database records with 2.69 bit/min from PostgreSQL, managed by a Python-Flask application, over the internet. Third, we leak secrets with 49.14 bit/min locally from ZRAM-compressed pages on Linux. Fourth, we leak internal heap pointers from the V8 engine within the Google Chrome browser on a system using ZRAM. This highlights the importance of re-evaluating the use of compression on sensitive data even if the application is only reachable via a remote interface