Microarchitectural attacks continue to evolve, uncovering new exploitation vectors in modern processors. From a defensive perspective, assessing a system's susceptibility to such attacks remains challenging. Developing functional attack implementations is labor-intensive, requires deep microarchitectural expertise, and is highly sensitive to execution environments. Consequently, existing attacks often lack portability, limiting systematic and scalable vulnerability assessment. Recent advances in large language models (LLMs) suggest a potential avenue for lowering these barriers. However, it remains unclear whether LLMs can reliably generate functionally correct microarchitectural attack code suitable for rigorous vulnerability testing. In this work, we present uGen, the first LLM-driven framework for automated microarchitectural attack code generation. A key challenge we address is identifying attack-specific knowledge gaps in LLMs. Through a systematic study of state-of-the-art models (GPT, Claude, and Qwen3), we find that LLMs frequently misgenerate or misplace critical attack primitives. Guided by this analysis, uGen employs a retrieval-augmented, multi-agent design that injects missing domain knowledge to synthesize functionally correct microarchitectural attack PoCs tailored to defender requirements. We evaluate uGen on cache-based and speculative-execution attacks across diverse set of microarchitectures, vulnerable functions, and LLM platforms. In the deployment stage, uGen achieves up to 100
This paper introduces BeaCon, a novel tool for the automated generation of adjustable container security policies. Unlike prior approaches, BeaCon leverages dynamic analysis to simulate realistic environments, uncovering container execution paths that may remain hidden during the profiling phase. To address the challenge of exploring vast profiling spaces, we employ efficient heuristics to reveal additional system events with minimal effort. In addition, BeaCon incorporates a security and functionality scoring mechanism to prioritize system calls and capabilities based on their impact on the host OS kernel’s security and the functionality of containerized applications. By integrating these scores, BeaCon achieves a customized balance between security and functionality, enabling cloud providers to enforce security measures while maintaining tenant availability. We implemented a prototype of BeaCon using eBPF kernel technology and conducted extensive evaluations. Results from the top 15 containers, which revealed significant improvements, demonstrate that BeaCon identifies an average of 16.5% additional syscalls by applying diverse environments. Furthermore, we evaluated its effectiveness in mitigating risks associated with 45 known vulnerabilities (e.g., CVEs), showcasing its potential to significantly enhance container security. Additionally, we performed proof-of-concept demonstrations for two well-known security vulnerabilities, showing that BeaCon successfully reduces attack surface by blocking these exploits.
Implantable Medical Devices (IMDs), such as pacemakers, increasingly communicate wirelessly with external devices. To secure this wireless communication channel, a pairing process is needed to bootstrap a secret key between the devices. Previous work has proposed pairing approaches that often adopt a “seamless” design and render the pairing process imperceptible to patients. This lack of user perception can significantly compromise security and pose threats to patients. The objective of this work is to explore the use of highly perceptible vibrations for pairing with IMDs. We aim to propose a novel technique that leverages the natural randomness in human motor behavior as a shared source of entropy for pairing, potentially deployable to current IMD products. We develop a proof-of-concept to demonstrate our proposed technique. We build a wearable prototype designed for individuals to simulate being an IMD patient (we do not test on real patients to avoid potential risks), and devise signal processing algorithms that utilize accelerometer readings to facilitate secure pairing with an IMD. We thoroughly evaluate the accuracy, security, and usability of our technique in a lab study with 24 participants. Our proposed pairing technique achieves high pairing accuracy, with a zero false acceptance rate (indicating low risks from adversaries) and a false rejection rate of only 0.6% (suggesting that legitimate users will likely experience very few failures). Our approach also offers robust security, which passes the National Institute of Standards and Technology statistical tests (with all p-values > 0.01). Moreover, our technique has high usability, evidenced by an average System Usability Scale questionnaire score of 73.6 (surpassing the standard benchmark of 68 for "good usability") and insights gathered from the interviews. Furthermore, the entire pairing process can be efficiently completed within five seconds. Vibration can be used to realize secure, usable, and deployable pairing in the context of IMDs. Our method also exhibits advantages over previous approaches, e.g., lenient requirements on the sensing capabilities of IMDs and the synchronization between the IMD and the external device.
Network topology verification in Software-Defined Networks (SDN) poses a significant challenge, as vulnerabilities can allow attackers to deceive the controller and manipulate the data plane into incorrect topologies, thereby endangering the entire network's security. Current solutions fail to guarantee both security and efficiency in the verification process, often resulting in damaging user traffic. With the aim of solving joint objectives, in this paper, we introduce PathSafe, a novel tool constructed on top of the existing controller frameworks designed for secure path verification in SDN environments. It enables the verification of all available paths between two points in the network and ensures a secure process. Our approach requires a data plane component for real-time packet monitoring at line speed and a control plane verification step. Our research demonstrates that PathSafe effectively mitigates security risks in compromised switches and host scenarios. Alongside a theoretical exploration of this challenge, we present a proof of concept implemented in P4, a common language for programmable data planes. Results obtained in Mininet underscore the practical applicability of PathSafe that, compared to alternatives, reduces overhead in the verification process while maintaining a limited execution time.
Serverless computing has rapidly emerged as a prominent cloud paradigm, enabling developers to focus solely on application logic without the burden of managing servers or underlying infrastructure. Public serverless repositories have become key to accelerating the development of serverless applications. However, their growing popularity makes them attractive targets for adversaries. Despite this, the security posture of these repositories remains largely unexplored, exposing developers and organizations to potential risks. In this paper, we present the first comprehensive analysis of the security landscape of serverless components hosted in public repositories. We analyse 2,758 serverless components from five widely used public repositories popular among developers and enterprises, and 125,936 Infrastructure as Code (IaC) templates across three widely used IaC frameworks. Our analysis reveals systemic vulnerabilities including outdated software packages, misuse of sensitive parameters, exploitable deployment configurations, susceptibility to typo-squatting attacks and opportunities to embed malicious behaviour within compressed serverless components. Finally, we provide practical recommendations to mitigate these threats.
Implantable Medical Devices (IMDs) are widely deployed today and often use wireless communication. Establishing a secure communication channel to these devices is challenging in practice. To address this issue, researchers have proposed IMD key exchange protocols, particularly ones that leverage an Out-Of-Band (OOB) channel such as audio, vibration and physiological signals. While these solutions have advantages over traditional key exchange, they are often proposed in an ad-hoc manner and lack a systematic evaluation of their security, usability and deployability properties. In this paper, we provide an in-depth analysis of existing OOB-based solutions for IMDs and, based on our findings, propose a novel IMD key exchange protocol that includes a new class of OOB channel based on human bodily motions. We implement prototypes and validate our designs through a user study (N = 24). The results demonstrate the feasibility of our approach and its unique features, establishing a new direction in the context of IMD security.
The computing continuum combines computational resources and services from edge to cloud, promising enhanced efficiency and resilience with respect to the traditional siloed-based approach. This study presents the REAR (Resource Advertisement and Reservation) protocol, which tackles the complexities of managing resources within this continuum. REAR establishes standardized interfaces to enable interoperability, enhances resource allocation efficiency, and maintains security measures for workload execution. The paper details the protocol’s design, key components, operational workflows, and potential uses, contributing to the optimization of resource use across the computing continuum.
Distributed SDN (Software-Defined Networking) controllers have rapidly become an integral element ofWide Area Networks (WAN), particularly within SD-WAN, providing scalability and fault-tolerance for expansive network infrastructures. However, the architecture of these controllers introduces new potential attack surfaces that have thus far received inadequate attention. In response to these concerns, we introduce Ambusher , a testing tool designed to discover vulnerabilities within protocols used in distributed SDN controllers. Ambusher achieves this by leveraging protocol state fuzzing , which systematically finds attack scenarios based on an inferred state machine. Since learning states from a cluster is complicated, Ambusher proposes a novel methodology that extracts a single and relatively simple state machine, achieving efficient state-based fuzzing. Our evaluation of Ambusher , conducted on a real SD-WAN deployment spanning two campus networks and one enterprise network, illustrates its ability to uncover 6 potential vulnerabilities in the widely used distributed controller platform.
Developing and deploying resource-aware artificial intelligence (AI) models presents a compelling optimization challenge in edge computing and serverless domains. Current research focuses mainly on scalable training and inference that uses serverless frameworks to optimize operational costs. However, these approaches often overlook the challenges of heterogeneous-aware training and overall cost optimization, including computing, memory, and communication. Our work introduces a framework for a heterogeneous edge environment, focusing on the model and key performance metrics like accuracy, floating-point operations per second, number of parameters, and latency. This framework enables distributed training within a serverless architecture. Further, it explores the construction of machine learning models from existing serverless functions using a lookup table while estimating AI model training in edge, cloud, or hybrid settings. We test the framework on object detection tasks in AI model development and deployment by using serverless operations.
The cloud computing landscape has evolved significantly in recent years, embracing various sandboxes to meet the diverse demands of modern cloud applications. These sandboxes encompass container-based technologies like Docker and gVisor, microVM-based solutions like Firecracker, and security-centric sandboxes relying on Trusted Execution Environments (TEEs) such as Intel SGX and AMD SEV. However, the practice of placing multiple tenants on shared physical hardware raises security and privacy concerns, most notably side-channel attacks. In this paper, we investigate the possibility of fingerprinting containers through CPU frequency reporting sensors in Intel and AMD CPUs. One key enabler of our attack is that the current CPU frequency information can be accessed by user-space attackers. We demonstrate that Docker images exhibit a unique frequency signature, enabling the distinction of different containers with up to 84.5% accuracy even when multiple containers are running simultaneously in different cores. Additionally, we assess the effectiveness of our attack when performed against several sandboxes deployed in cloud environments, including Google's gVisor, AWS' Firecracker, and TEE-based platforms like Gramine (utilizing Intel SGX) and AMD SEV. Our empirical results show that these attacks can also be carried out successfully against all of these sandboxes in less than 40 seconds, with an accuracy of over 70% in all cases. Finally, we propose a noise injection-based countermeasure to mitigate the proposed attack on cloud environments.
Cryptographic primitives are of paramount importance to guarantee security properties in communication networks. The associated computational complexity of cryptography standards makes it prohibitive to execute these primitives at line rate in the network core. Existing implementations of cryptographic MAC algorithms in P4 for programmable switches impose a severe performance penalty due to packet recirculation, which may not be tolerable at those network speeds. In this paper, we propose the first data plane design in P4 of the Chaskey algorithm, a widely used secure and lightweight cryptographic MAC algorithm, tailored for the PISA switch architecture. Our P4CHASKEY is the first solution to compute MACs using 128-bit keys without packet recirculation, guaranteeing line rate Terabit speeds. As state-of-the-art solutions require recirculations for the same key size (reducing throughput performance) or offer weaker security (smaller keys), P4CHASKEY is now, to our knowledge, the most efficient MAC design for the target switch architecture.
During the last few years, Implantable Medical Devices (IMDs) have evolved considerably. IMD manufacturers are now starting to rely on standard wireless technologies for connectivity. Moreover, there is an evolution towards open systems where the IMD can be remotely monitored or reconfigured through personal commercial-off-the-shelf devices such as smartphones or tablets. Nevertheless, a major problem that still remains unsolved today is the secure establishment of cryptographic keys between the IMD and such personal devices. Researchers have already proposed various solutions, most notably by relying on an additional external device. Unfortunately, these proposed approaches are either insecure, difficult to realise in practice, or are unsuitable for the latest generation of IMDs. Motivated by this, we present HAT, a secure and practical solution to provide fine-grained and dynamic access control for the next generation of IMDs, while offering full control and transparency to the patient. The main idea behind HAT is to shift the access control responsibilities from the IMD to an external device under the user's control, such as a smartphone, acting as the IMD's Key Distribution Center. We show that HAT only introduces minimal energy and memory overhead and formally prove its security using Verifpal.
In the rapidly evolving field of serverless computing, efficient function scheduling and resource scaling are critical for optimizing performance and cost. This paper presents a comprehensive review of the application of Deep Reinforcement Learning (DRL) techniques in these areas. We begin by providing an overview of serverless computing, highlighting its benefits and challenges, with a particular focus on function scheduling and resource scaling. We then delve into the principles of deep reinforcement learning (DRL) and its potential for addressing these challenges. A systematic review of recent studies applying DRL to serverless computing is presented, covering various algorithms, models, and performances. Our analysis reveals that DRL, with its ability to learn and adapt from an environment, shows promising results in improving the efficiency of function scheduling and resource scaling in serverless computing. However, several challenges remain, including the need for more realistic simulation environments, handling of cold starts, and the trade-off between learning time and scheduling performance. We conclude by discussing potential future directions for this research area, emphasizing the need for more robust DRL models, better benchmarking methods, and the exploration of multi-agent reinforcement learning for more complex serverless architectures. This review serves as a valuable resource for researchers and practitioners aiming to understand and advance the application of DRL in serverless computing.
Cloud-native services face unique cybersecurity challenges due to their distributed infrastructure. They are susceptible to various threats like malware, DDoS attacks, and Man-in-the-Middle (MITM) attacks. Additionally, these services often process sensitive data that must be protected from unauthorized access. On top of that, the dynamic and scalable nature of cloud-native services makes it difficult to maintain consistent security, as deploying new instances and infrastructure introduces new vulnerabilities. To address these challenges, efficient security solutions are needed to mitigate potential threats while aligning with the characteristics of cloud-native services. Despite the abundance of works focusing on security aspects in the cloud, there has been a notable lack of research that is focused on the security of cloud-native services. To address this gap, this work is the first survey that is dedicated to exploring security in cloud-native services. This work aims to provide a comprehensive investigation of the aspects, features, and solutions that are associated with security in cloud-native services. It serves as a uniquely structured mapping study that maps the key aspects to the corresponding features, and these features to numerous contemporary solutions. Furthermore, it includes the identification of various candidate open-source technologies that are capable of supporting the realization of each explored solution. Finally, it showcases how these solutions can work together in order to establish each corresponding feature. The insights and findings of this work can be used by cybersecurity professionals, such as developers and researchers, to enhance the security of cloud-native services.
Software-defined wide area network (SD-WAN) has emerged as a new paradigm for steering a large-scale network flexibly by adopting distributed software-defined network (SDN) controllers. The key to building a logically centralized but physically distributed control-plane is running diverse cluster management protocols to achieve consistency through an exchange of control traffic. Meanwhile, we observe that the control traffic exposes unique time-series patterns and directional relationships due to the operational structure even though the traffic is encrypted, and this pattern can disclose confidential information such as control-plane topology and protocol dependencies, which can be exploited for severe attacks. With this insight, we propose a new SD-WAN fingerprinting system, called Heimdallr. It analyzes periodical and operational patterns of SD-WAN cluster management protocols and the context of flow directions from the collected control traffic utilizing a deep learning-based approach, so that it can classify the cluster management protocols automatically from miscellaneous control traffic datasets. Our evaluation, which is performed in a realistic SD-WAN environment consisting of geographically distant three campus networks and one enterprise network shows that Heimdallr can classify SD-WAN control traffic with ≥ 93%, identify individual protocols with ≥ 80% macro F-1 scores, and finally can infer control-plane topology with ≥ 70% similarity.
Hardware tokens are increasingly used to support second-factor and passwordless authentication schemes. While these devices improve security over weaker factors like passwords, they suffer from a number of security and practical issues. We present the design and implementation of Symbolon, a system that allows users to authenticate to an online service in a secure and flexible manner by using multiple personal devices (e.g., their smartphone and smart watch) together, in place of a password. The core idea behind Symbolon is to let users authenticate only if they carry a sufficient number of their personal devices and give explicit consent. We use threshold cryptography at the client side to protect against strong adversaries while overcoming the limitations of multi-factor authentication in terms of flexibility. Symbolon is compatible with FIDO servers, but improves the client-side experience compared to FIDO in terms of security, privacy, and user control. We design Symbolon such that the user can (i) authenticate using a flexible selection of devices, which we call “authenticators”; (ii) define fine-grained threshold policies that enforce user consent without involving or modifying online services; and (iii) add or revoke authenticators without needing to generate new cryptographic keys or manually (un)register them with online services. Finally, we present a detailed design and analyse the security, privacy and practical properties of Symbolon; this includes a formal proof using ProVerif to show the required security properties are satisfied.
Implantable medical devices, sensors and wearables are widely deployed today. However, establishing a secure wireless communication channel to these devices is a major challenge, amongst others due to the constraints on energy consumption and the need to obtain immediate access in emergencies. To address this issue, researchers have proposed various key agreement protocols based on the measurement of physiological signals such as a person’s heart signal. At the core of such protocols are fuzzy cryptographic primitives that allow to agree on a shared secret based on several simultaneous, noisy measurements of the same signal. So far, although many fuzzy primitives have been proposed, there is no comprehensive evaluation and comparison yet of the overhead that such methods incur on resource-constrained embedded devices. In this paper, we study the feasibility of six types of fuzzy cryptographic primitives on embedded devices for 128-bit key agreement. We configure several variants for each fuzzy primitive under different parameter selections and mismatch rates of the physiological signal measurements on an MSP430 microcontroller, and then measure and compare their energy consumption and communication overhead. The most efficient constructions consume between 0.021 mJ and 0.198 mJ for the transmitter and between 0.029 mJ and 0.380 mJ for the receiver under different mismatch rates. Subsequently, we modify the best performing methods so that they run in constant time to protect against timing side-channel attacks, and observe that these changes only minimally affect resource consumption. Finally, we provide open-source implementations and energy consumption data of each fuzzy primitive as a reference for real-world designs.
ment mechanisms, Quality of Service (QoS) and Quality of Experience (QoE) management, network security and many other tasks.1 On the industrial side, AI is slowly complementing traditional networking approaches worldwide: Small Medium Enterprises (SMEs) and start-ups are developing AI solutions to deal with specific use cases, traditional networking vendors are evolving their products to support AI tools, and major cloud/software providers are adapting AI tools to be used in the networking domain. This is also the case for telephone service providers (telcos), which are exploring the application of AI algorithms through internal research and innovation (R&I) projects. For instance, our group is working on AI-based approaches in many use-cases focusing on realistic environments and applications (for example, Kattadige et al.2 and Perino property inference or data reconstruction attacks, and adversarial learning, can reveal different aspects of the data used (for example, which specific users’ data were used for model training), the values of their data attributes, and even user patterns such as their mobility or browsing behavior. Therefore, the use of AI techniques can impact user privacy more strongly than traditional data analysis methods since AI models can distill information from multiple data sources and infer rich patterns regarding et al.5), partially in collaboration with other European partners in the context of European R&I actions.b What about my privacy? AI models and tools are potentially vulnerable, and their usage introduces new attack vectors for telco environments. For instance, membership and
In this article we review the current serverless architectures, abstract and categorize their founding principles, and provide an in-depth security analysis. In particular, we: show the security shortcomings of the analyzed serverless architectural paradigms; point to possible countermeasures; and, highlight several research directions for practitioners, Industry, and Academia.