arXiv subjects

Srdjan Capkun

Publications and source records attributed to Srdjan Capkun.

At least 19 recordsLinked to original sources

arXiv2026

MOSAIC: Masked Outsourcing of Secure AI Computations

We address the challenge of securely and efficiently outsourcing AI computations from a trusted but computationally weak client to an untrusted but powerful server, in the setting where the client holds both the input and the model, and the server must learn neither. We present MOSAIC, whose core is a novel matrix-multiplication masking protocol that scales to far larger matrices than prior work, enabling the safe outsourcing of modern workloads such as large transformer inference. By introducing small amounts of noise to the multiplication result and thereby relaxing correctness, MOSAIC achieves optimal asymptotic client overhead and concrete runtimes orders of magnitude faster than prior work. Its security reduces to the decisional LWE and LPN assumptions. Because this noise accumulates across the many layers of a transformer, a key technical challenge is bounding error growth; MOSAIC addresses this with an error-scaling mechanism based on random Hadamard rotations. On large 70B transformer models, MOSAIC's perplexity is comparable to popular quantization approaches and even matches full-precision BF16 inference on HumanEval. Finally, we present an end-to-end implementation showing how ideas like MOSAIC can promise a path towards large-scale confidential AI in modern data centers. Non-confidential inference is already distributed across phase (prefill/decode), layer, and time to maximize utilization of heterogeneous hardware, using RDMA-like networking to move activations, cached KV values, and weights across nodes. MOSAIC enables scaling of confidential compute by keeping the trusted computing base (TCB) small and outsourcing the bulk of the AI computation to untrusted accelerators.

cs.CR↗

arXiv2026

Gyokuro: Source-assisted Private Membership Testing using Trusted Execution Environments

Private Membership Testing (PMT) protocols enable clients to verify whether a certain data item is included in a database without revealing the item to the database operator or other external parties. This paper examines Source-assisted PMT (SPMT), in which clients leverage compact data source-provided information issued when the data item is first submitted to the database. SPMT is relevant in applications such as certificate transparency and supply-chain auditing; yet, designing an approach that is efficient, scalable, and privacy-preserving remains a challenge. This work presents Gyokuro, which takes a different approach to conventional membership testing schemes. Instead of requesting the server to produce a proof attesting that a certain data item exists in the database, we leverage Trusted Execution Environments (TEEs) to produce proofs demonstrating that the server has made enough progress to add the data item to the database. With the help of existing monitoring services, clients can infer that no items have been removed from the database. This allows Gyokuro to provide strong privacy guaranties and achieve high efficiency, as a client's membership testing query does not include any information regarding their interests, and eliminates the need for complex and inefficient protection mechanisms. Additionally, this approach enables membership testing on large-scale databases, since the communication and computation required are independent of the database size. Our evaluations show practical feasibility, achieving 7 ms membership testing latency and throughput of around 1400 requests/sec/core.

cs.CR↗

arXiv2026

5Gone: Uplink Overshadowing Attacks in 5G-SA

5G presents numerous advantages compared to previous generations: improved throughput, lower latency, and improved privacy protection for subscribers. Attacks against 5G standalone (SA) commonly use fake base stations (FBS), which need to operate at a very high output power level to lure victim phones to connect to them and are thus highly detectable. In this paper, we introduce 5Gone, a powerful software-defined radio (SDR)-based uplink overshadowing attack method against 5G-SA. 5Gone exploits deficiencies in the 3GPP standard to perform surgical, covert denial-of-service, privacy, and downgrade attacks. Uplink overshadowing means that an attacker is transmitting at exactly the same time and frequency as the victim UE, but with a slightly higher output power. 5Gone runs on a COTS x86 computer without any need for dedicated hardware acceleration and can overshadow commercial 100 MHz cells with an E2E latency of less than 500$\mu$s, which up to now has not been possible with any software-based UE implementation. We demonstrate that 5Gone is highly scalable, even when many UEs are connecting in parallel, and finally evaluate the attacks end-to-end against 7 phone models and three different chipset vendors both in our lab and in the real-world on public gNodeBs.

cs.CR↗

arXiv2025

AEX-NStep: Probabilistic Interrupt Counting Attacks on Intel SGX

To mitigate interrupt-based stepping attacks (notably using SGX-Step), Intel introduced AEX-Notify, an ISA extension to Intel SGX that aims to prevent deterministic single-stepping. In this work, we introduce AEX-NStep, the first interrupt counting attack on AEX-Notify-enabled Enclaves. We show that deterministic single-stepping is not required for interrupt counting attacks to be practical and that, therefore, AEX-Notify does not entirely prevent such attacks. We specifically show that one of AEX-Notify's security guarantees, obfuscated forward progress, does not hold, and we introduce two new probabilistic interrupt counting attacks. We use these attacks to construct a practical ECDSA key leakage attack on an AEX-Notify-enabled SGX enclave. Our results extend the original security analysis of AEX-Notify and inform the design of future mitigations.

cs.CR↗

arXiv2025

URL Inspection Tasks: Helping Users Detect Phishing Links in Emails

The most widespread type of phishing attack involves email messages with links pointing to malicious content. Despite user training and the use of detection techniques, these attacks are still highly effective. Recent studies show that it is user inattentiveness, rather than lack of education, that is one of the key factors in successful phishing attacks. To this end, we develop a novel phishing defense mechanism based on URL inspection tasks: small challenges (loosely inspired by CAPTCHAs) that, to be solved, require users to interact with, and understand, the basic URL structure. We implemented and evaluated three tasks that act as ``barriers'' to visiting the website: (1) correct click-selection from a list of URLs, (2) mouse-based highlighting of the domain-name URL component, and (3) re-typing the domain-name. These tasks follow best practices in security interfaces and warning design. We assessed the efficacy of these tasks through an extensive on-line user study with 2,673 participants from three different cultures, native languages, and alphabets. Results show that these tasks significantly decrease the rate of successful phishing attempts, compared to the baseline case. Results also showed the highest efficacy for difficult URLs, such as typo-squats, with which participants struggled the most. This highlights the importance of (1) slowing down users while focusing their attention and (2) helping them understand the URL structure (especially, the domain-name component thereof) and matching it to their intent.

cs.CR↗

arXiv2024

Breaking Bad: How Compilers Break Constant-Time Implementations

The implementations of most hardened cryptographic libraries use defensive programming techniques for side-channel resistance. These techniques are usually specified as guidelines to developers on specific code patterns to use or avoid. Examples include performing arithmetic operations to choose between two variables instead of executing a secret-dependent branch. However, such techniques are only meaningful if they persist across compilation. In this paper, we investigate how optimizations used by modern compilers break the protections introduced by defensive programming techniques. Specifically, how compilers break high-level constant-time implementations used to mitigate timing side-channel attacks. We run a large-scale experiment to see if such compiler-induced issues manifest in state-of-the-art cryptographic libraries. We develop a tool that can profile virtually any architecture, and we use it to run trace-based dynamic analysis on 44,604 different targets. Particularly, we focus on the most widely deployed cryptographic libraries, which aim to provide side-channel resistance. We are able to evaluate whether their claims hold across various CPU architectures, including x86-64, x86-i386, armv7, aarch64, RISC-V, and MIPS-32. Our large-scale study reveals that several compiler-induced secret-dependent operations occur within some of the most highly regarded hardened cryptographic libraries. To the best of our knowledge, such findings represent the first time these issues have been observed in the wild. One of the key takeaways of this paper is that the state-of-the-art defensive programming techniques employed for side-channel resistance are still inadequate, incomplete, and bound to fail when paired with the optimizations that compilers continuously introduce.

cs.CR↗

arXiv2024

Content, Nudges and Incentives: A Study on the Effectiveness and Perception of Embedded Phishing Training

A common form of phishing training in organizations is the use of simulated phishing emails to test employees' susceptibility to phishing attacks, and the immediate delivery of training material to those who fail the test. This widespread practice is dubbed embedded training; however, its effectiveness in decreasing the likelihood of employees falling for phishing again in the future is questioned by the contradictory findings of several recent field studies. We investigate embedded phishing training in three aspects. First, we observe that the practice incorporates different components -- knowledge gains from its content, nudges and reminders from the test itself, and the deterrent effect of potential consequences -- our goal is to study which ones are more effective, if any. Second, we explore two potential improvements to training, namely its timing and the use of incentives. Third, we analyze employees' reception and perception of the practice. For this, we conducted a large-scale mixed-methods (quantitative and qualitative) study on the employees of a partner company. Our study contributes several novel findings on the training practice: in particular, its effectiveness comes from its nudging effect, i.e., the periodic reminder of the threat rather than from its content, which is rarely consumed by employees due to lack of time and perceived usefulness. Further, delaying training to ease time pressure is as effective as currently established practices, while rewards do not improve secure behavior. Finally, some of our results support previous findings with increased ecological validity, e.g., that phishing is an attention problem, rather than a knowledge one, even for the most susceptible employees, and thus enforcing training does not help.

cs.CR↗

arXiv2024

PayOff: A Regulated Central Bank Digital Currency with Private Offline Payments

The European Central Bank is preparing for the potential issuance of a central bank digital currency (CBDC), called the digital euro. A recent regulatory proposal by the European Commission defines several requirements for the digital euro, such as support for both online and offline payments. Offline payments are expected to enable cash-like privacy, local payment settlement, and the enforcement of holding limits. While other central banks have expressed similar desired functionality, achieving such offline payments poses a novel technical challenge. We observe that none of the existing research solutions, including offline E-cash schemes, are fully compliant. Proposed solutions based on secure elements offer no guarantees in case of compromise and can therefore lead to significant payment fraud. The main contribution of this paper is PayOff, a novel CBDC design motivated by the digital euro regulation, which focuses on offline payments. We analyze the security implications of local payment settlement and identify new security objectives. PayOff protects user privacy, supports complex regulations such as holding limits, and implements safeguards to increase robustness against secure element failure. Our analysis shows that PayOff provides strong privacy and identifies residual leakages that may arise in real-world deployments. Our evaluation shows that offline payments can be fast and that the central bank can handle high payment loads with moderate computing resources. However, the main limitation of PayOff is that offline payment messages and storage requirements grow in the number of payments that the sender makes or receives without going online in between.

cs.CR↗

arXiv2024

Devlore: Device Interrupt Protection for Confidential VMs

Modern confidential computing executes sensitive computation in an abstraction called confidential VMs and protects from the hypervisor, host OS, and other co-resident VMs. It has been shown that an attacker can inject malicious interrupts to break the confidentiality and integrity of confidential VMs. We present Devlore, a device interrupt isolation mechanism that protects confidential VMs from interrupt manipulation attacks. Our design employs a delegate-but-check strategy by offloading interrupt management to the hypervisor, but adds correctness checks in the trusted software. We prototype our design on Arm Confidential Computing Architecture (CCA). We evaluate it on Arm FVP to demonstrate four diverse devices attached to confidential VMs and report costs on a Rock5b board. Our case studies show the feasibility of real-world use cases and that Devlore incurs minimal overheads of 0.06% for typical integrated GPU applications.

cs.CR↗

arXiv2024

Aster: Fixing the Android TEE Ecosystem with Arm CCA

The Android ecosystem relies on either TrustZone (e.g., OP-TEE, QTEE, Trusty) or trusted hypervisors (pKVM, Gunyah) to isolate security-sensitive services from malicious apps and Android bugs. TrustZone allows any secure world code to access the normal world that runs Android. Similarly, a trusted hypervisor has full access to Android running in one VM and security services in other VMs. In this paper, we motivate the need for mutual isolation, wherein Android, hypervisors, and the secure world are isolated from each other. Then, we propose a sandboxed service abstraction, such that a sandboxed execution cannot access any other sandbox, Android, hypervisor, or secure world memory. We present Aster which achieves these goals while ensuring that sandboxed execution can still communicate with Android to get inputs and provide outputs securely. Our main insight is to leverage the hardware isolation offered by Arm Confidential Computing Architecture (CCA). However, since CCA does not satisfy our sandboxing and mutual isolation requirements, Aster repurposes its hardware enforcement to meet its goals while addressing challenges such as secure interfaces, virtio, and protection against interrupts. We implement Aster to demonstrate its feasibility and assess its compatibility. We take three case studies, including one currently deployed on Android phones and insufficiently secured using a trusted hypervisor, to demonstrate that they can be protected by Aster.

cs.CR↗

arXiv2023

Time for Change: How Clocks Break UWB Secure Ranging

Due to its suitability for wireless ranging, Ultra-Wide Band (UWB) has gained traction over the past years. UWB chips have been integrated into consumer electronics and considered for security-relevant use cases, such as access control or contactless payments. However, several publications in the recent past have shown that it is difficult to protect the integrity of instance measurements on the physical layer. In this paper, we identify transceiver clock imperfections as a new, important parameter that has been widely ignored so far. We present Mix-Down and Stretch-and-Advance, two novel attacks against the current (IEEE 802.15.4z) and the upcoming (IEEE 802.15.4ab) UWB standard, respectively. We demonstrate Mix-Down on commercial chips and achieve distance reduction from 10 m to 0 m. For the Stretch-and-Advance attack, we show analytically that the current proposal of IEEE 802.15.4ab allows reductions of over 90 m. In order to prevent the attack, we propose and analyze an effective countermeasure.

cs.CR↗

arXiv2022

It's TEEtime: A New Architecture Bringing Sovereignty to Smartphones

Modern smartphones are complex systems in which control over phone resources is exercised by phone manufacturers, OS vendors, and users. These stakeholders have diverse and often competing interests. Barring some exceptions, users entrust their security and privacy to OS vendors (Android and iOS) and need to accept their constraints. Manufacturers protect their firmware and peripherals from the OS by executing in the highest privilege and leveraging dedicated CPUs and TEEs. OS vendors need to trust the highest privileged code deployed by manufacturers. This division of control over the phone is not ideal for OS vendors and is even more disadvantageous for the users. Users are generally limited in what applications they can install on their devices, in the privacy model and trust assumptions of the existing applications, and in the functionalities that applications can have. We propose TEEtime, a new smartphone architecture based on trusted execution allowing to balance the control different stakeholders exert over phones. More leveled control over the phone means that no stakeholder is more privileged than the others. In particular, TEEtime makes users sovereign over their phones: It enables them to install sensitive applications in isolated domains with protected access to selected peripherals alongside an OS. TEEtime achieves this while maintaining compatibility with the existing smartphone ecosystem and without relying on virtualization; it only assumes trust in a phone's firmware. TEEtime is the first TEE architecture that allows isolated execution domains to gain protected and direct access to peripherals. TEEtime is based on Armv8-A and achieves peripheral isolation using a novel mechanism based on memory and interrupt controller protection. We demonstrate the feasibility of our design by implementing a prototype of TEEtime, and by running exemplary sensitive applications.

cs.CR↗

arXiv2022

Empowering Data Centers for Next Generation Trusted Computing

Modern data centers have grown beyond CPU nodes to provide domain-specific accelerators such as GPUs and FPGAs to their customers. From a security standpoint, cloud customers want to protect their data. They are willing to pay additional costs for trusted execution environments such as enclaves provided by Intel SGX and AMD SEV. Unfortunately, the customers have to make a critical choice -- either use domain-specific accelerators for speed or use CPU-based confidential computing solutions. To bridge this gap, we aim to enable data-center scale confidential computing that expands across CPUs and accelerators. We argue that having wide-scale TEE-support for accelerators presents a technically easier solution, but is far away from being a reality. Instead, our hybrid design provides enclaved execution guarantees for computation distributed over multiple CPU nodes and devices with/without TEE support. Our solution scales gracefully in two dimensions -- it can handle a large number of heterogeneous nodes and it can accommodate TEE-enabled devices as and when they are available in the future. We observe marginal overheads of $0.42$--$8\%$ on real-world AI data center workloads that are independent of the number of nodes in the data center. We add custom TEE support to two accelerators (AI and storage) and integrate it into our solution, thus demonstrating that it can cater to future TEE devices.

cs.CR↗

arXiv2022

SoK: Hardware-supported Trusted Execution Environments

The growing complexity of modern computing platforms and the need for strong isolation protections among their software components has led to the increased adoption of Trusted Execution Environments (TEEs). While several commercial and academic TEE architectures have emerged in recent times, they remain hard to compare and contrast. More generally, existing TEEs have not been subject to a holistic systematization to understand the available design alternatives for various aspects of TEE design and their corresponding pros-and-cons. Therefore, in this work, we analyze the design of existing TEEs and systematize the mechanisms that TEEs implement to achieve their security goals, namely, verifiable launch, run-time isolation, trusted IO, and secure storage. More specifically, we analyze the typical architectural building blocks underlying TEE solutions, design alternatives for each of these components and the trade-offs that they entail. We focus on hardware-assisted TEEs and cover a wide range of TEE proposals from academia and the industry. Our analysis shows that although TEEs are diverse in terms of their goals, usage models, and instruction set architectures, they all share many common building blocks in terms of their design.

cs.CR↗

arXiv2021

Phishing in Organizations: Findings from a Large-Scale and Long-Term Study

In this paper, we present findings from a large-scale and long-term phishing experiment that we conducted in collaboration with a partner company. Our experiment ran for 15 months during which time more than 14,000 study participants (employees of the company) received different simulated phishing emails in their normal working context. We also deployed a reporting button to the company's email client which allowed the participants to report suspicious emails they received. We measured click rates for phishing emails, dangerous actions such as submitting credentials, and reported suspicious emails. The results of our experiment provide three types of contributions. First, some of our findings support previous literature with improved ecological validity. One example of such results is good effectiveness of warnings on emails. Second, some of our results contradict prior literature and common industry practices. Surprisingly, we find that embedded training during simulated phishing exercises, as commonly deployed in the industry today, does not make employees more resilient to phishing, but instead it can have unexpected side effects that can make employees even more susceptible to phishing. And third, we report new findings. In particular, we are the first to demonstrate that using the employees as a collective phishing detection mechanism is practical in large organizations. Our results show that such crowd-sourcing allows fast detection of new phishing campaigns, the operational load for the organization is acceptable, and the employees remain active over long periods of time.

cs.CR↗

arXiv2021

Ghost Peak: Practical Distance Reduction Attacks Against HRP UWB Ranging

We present the first over-the-air attack on IEEE 802.15.4z High-Rate Pulse Repetition Frequency (HRP) Ultra-WideBand (UWB) distance measurement systems. Specifically, we demonstrate a practical distance reduction attack against pairs of Apple U1 chips (embedded in iPhones and AirTags), as well as against U1 chips inter-operating with NXP and Qorvo UWB chips. These chips have been deployed in a wide range of phones and cars to secure car entry and start and are projected for secure contactless payments, home locks, and contact tracing systems. Our attack operates without any knowledge of cryptographic material, results in distance reductions from 12m (actual distance) to 0m (spoofed distance) with attack success probabilities of up to 4%, and requires only an inexpensive (USD 65) off-the-shelf device. Access control can only tolerate sub-second latencies to not inconvenience the user, leaving little margin to perform time-consuming verifications. These distance reductions bring into question the use of UWB HRP in security-critical applications.

cs.CR↗

arXiv2021

LTrack: Stealthy Tracking of Mobile Phones in LTE

We introduce LTrack, a new tracking attack on LTE that allows an attacker to stealthily extract user devices' locations and permanent identifiers (IMSI). To remain stealthy, the localization of devices in LTrack is fully passive, relying on our new uplink/downlink sniffer. Our sniffer records both the times of arrival of LTE messages and the contents of the Timing Advance Commands, based on which LTrack calculates locations. LTrack is the first to show the feasibility of a passive localization in LTE through implementation on software-defined radio. Passive localization attacks reveal a user's location traces but can at best link these traces to a device's pseudonymous temporary identifier (TMSI), making tracking in dense areas or over a long time-period challenging. LTrack overcomes this challenge by introducing and implementing a new type of IMSI Catcher named IMSI Extractor. It extracts a device's IMSI and binds it to its current TMSI. Instead of relying on fake base stations like existing IMSI Catchers, which are detectable due to their continuous transmission, IMSI Extractor relies on our uplink/downlink sniffer enhanced with surgical message overshadowing. This makes our IMSI Extractor the stealthiest IMSI Catcher to date. We evaluate LTrack through a series of experiments and show that in line-of-sight conditions, the attacker can estimate the location of a phone with less than 6m error in 90% of the cases. We successfully tested our IMSI Extractor against a set of 17 modern smartphones connected to our industry-grade LTE testbed. We further validated our uplink/downlink sniffer and IMSI Extractor in a test facility of an operator.

cs.CR↗

arXiv2021

AdaptOver: Adaptive Overshadowing Attacks in Cellular Networks

In cellular networks, attacks on the communication link between a mobile device and the core network significantly impact privacy and availability. Up until now, fake base stations have been required to execute such attacks. Since they require a continuously high output power to attract victims, they are limited in range and can be easily detected both by operators and dedicated apps on users' smartphones. This paper introduces AdaptOver - a MITM attack system designed for cellular networks, specifically for LTE and 5G-NSA. AdaptOver allows an adversary to decode, overshadow (replace) and inject arbitrary messages over the air in either direction between the network and the mobile device. Using overshadowing, AdaptOver can cause a persistent ($\geq$ 12h) DoS or a privacy leak by triggering a UE to transmit its persistent identifier (IMSI) in plain text. These attacks can be launched against all users within a cell or specifically target a victim based on its phone number. We implement AdaptOver using a software-defined radio and a low-cost amplification setup. We demonstrate the effects and practicality of the attacks on a live operational LTE and 5G-NSA network with a wide range of smartphones. Our experiments show that AdaptOver can launch an attack on a victim more than 3.8km away from the attacker. Given its practicability and efficiency, AdaptOver shows that existing countermeasures that are focused on fake base stations are no longer sufficient, marking a paradigm shift for designing security mechanisms in cellular networks.

cs.CR↗