As satellite systems become a greater part of critical infrastructure, they have become a significantly more appealing target for attacks. The availability of cheap off-the-shelf radio hardware has made signal spoofing and physical layer attacks more accessible than ever to a wide range of adversaries, from hobbyists to nation-state actors. Legacy systems are particularly vulnerable due to their lack of cryptographic security, and cannot be patched to support novel security measures. In this article, we use radio transmitter fingerprinting to authenticate satellite downlinks, using characteristics of the transmitter hardware expressed as impairments on the physical layer radio signal. Our SATIQ system employs a Siamese neural network and an autoencoder to extract an efficient encoding of message headers that preserves identifying information. We focus on high sample rate fingerprinting, making device fingerprints difficult to forge without similarly high sample rate transmitting hardware. We collected 10 290 000 messages from the Iridium satellite constellation at 25 MS/s, and demonstrate that the SATIQ model trained on this data maintains performance over time without retraining, and can be used on new transmitters with no impact on performance. We analyze the system's robustness against weather and signal factors, and demonstrate its effectiveness under attack, achieving an Equal Error Rate of 0.072 and ROC AUC of 0.960. We conclude that our techniques are useful for building fingerprinting systems that are effective at authenticating satellite communication, maintain performance over time and across satellite replacement, and provide robustness against spoofing and replay by raising the required budget for attacks.

Recent advances in large language models (LLMs) have enabled a new generation of autonomous agents that operate over sustained periods and manage sensitive resources on behalf of users. Trusted for their ability to act without direct oversight, such agents are increasingly considered in high-stakes domains including financial management, dispute resolution, and governance. Yet in practice, agents execute on infrastructure controlled by a host, who can tamper with models, inputs, or outputs, undermining any meaningful notion of autonomy. We address this gap by introducing VET (Verifiable Execution Traces), a formal framework that achieves host-independent authentication of agent outputs and takes a step toward host-independent autonomy. Central to VET is the Agent Identity Document (AID), which specifies an agent's configuration together with the proof systems required for verification. VET is compositional: it supports multiple proof mechanisms, including trusted hardware, succinct cryptographic proofs, and notarized TLS transcripts (Web Proofs). We implement VET for an API-based LLM agent and evaluate our instantiation on realistic workloads. We find that for today's black-box, secret-bearing API calls, Web Proofs appear to be the most practical choice, with overhead typically under 3× compared to direct API calls, while for public API calls, a lower-overhead TEE Proxy is often sufficient. As a case study, we deploy a verifiable trading agent that produces proofs for each decision and composes Web Proofs with a TEE Proxy. Our results demonstrate that practical, host-agnostic authentication is already possible with current technology, laying the foundation for future systems that achieve full host-independent autonomy.

Simulation tools are commonly used in the development and testing of new protocols or new networks. However, as satellite networks start to grow to encompass thousands of nodes, and as companies and space agencies begin to realize the interplanetary internet, existing satellite and network simulation tools have become impractical for use in this context. We therefore present the Deep Space Network Simulator (DSNS): a new network simulator with a focus on large-scale satellite networks. We demonstrate its improved capabilities compared to existing offerings, showcase its flexibility and extensibility through an implementation of existing protocols and the DTN simulation reference scenarios recommended by CCSDS, and evaluate its scalability, showing that it exceeds existing tools while providing better fidelity. DSNS provides concrete usefulness to both standards bodies and satellite operators, enabling fast iteration on protocol development and testing of parameters under highly realistic conditions. By removing roadblocks to research and innovation, we can accelerate the development of upcoming satellite networks and ensure that their communication is both fast and secure.

Motivated by the growing prevalence of increasingly advanced satellite jamming attacks, we introduce and systematically analyze protocol-aware jammers: the worst-case scenario that maximally exploits the protocol to deny service whilst remaining as difficult to detect as possible. This extends existing satellite jamming and anti-jamming literature, which to date considers only conventional jamming waveforms. We find that protocol-aware jammers are significantly more effective than conventional jammers against all major standardized satellite protocols, including when anti-jamming countermeasures in the form of interleaving and adaptive coding and modulation are employed. This performance is possible since current protocols have a cyclic and predictable nature. We assess the required capabilities in terms of synchronization, and showthat many of these performance gains can be realized even by completely desynchronized jammers. We experimentally evaluate protocol-aware strategies against both a hardware and software receiver. The results show that over 15 dB of performance gains over Gaussian jamming are possible against all tested satellite protocols. Furthermore, we find that the attack can be optimized in simulation and deployed against the hardware receiver without performance degradation. We conclude with a discussion of countermeasures, primarily at the protocol level, to improve the availability of these systems.

Due to the increasing threat of attacks on satellite systems, novel countermeasures have been developed to provide additional security. Among these, there has been a particular interest in transmitter fingerprinting, which authenticates transmitters by looking at characteristics expressed in the physical layer signal. These systems rely heavily upon statistical methods and machine learning, and are therefore vulnerable to a range of attacks. The severity of this threat in a fingerprinting context is currently not well understood. In this paper we evaluate a range of attacks against satellite fingerprinting, building on previous works by looking at attacks optimized to target the fingerprinting system for maximal impact. We design optimized jamming, dataset poisoning, and spoofing attacks, evaluating them in the real world against the SatIQ fingerprinting system designed to authenticate Iridium transmitters, and using a wireless channel emulator to achieve realistic channel conditions. We show that an optimized jamming signal can cause a 50 Finally, we show that a model trained to optimize spoofing attacks can also be used to detect spoofing and replay attacks, even when it has never seen the attacker's transmitter before. This technique works even when the training dataset includes only a single transmitter, enabling fingerprinting to be used to protect small constellations and even individual satellites, providing additional protection where it is needed the most.

Direct Sequence Spread Spectrum (DSSS) is used to simplify frequency management for constellations and for use of data relay satellites, to improve satellite mission availability against unintentional interference and protect space RF links against jamming, eavesdropping, and spoofing. Whilst current standards focus on cooperative Code Division Multiple Access (CDMA) DSSS methods, high-value government and military assets increasingly use cryptographic DSSS to improve security. Including cryptographic DSSS into future revisions of the ETSI standard is currently considered an option, but it has been found that cryptographic DSSS is significantly worse at multiple access than the currently standardized methods. In this context, the European Space Agency and Thales Alenia Space have studied a hybrid CDMA/cryptographic DSSS construction designed to simultaneously provide multiple-access and security. In this paper we perform the first systematic analysis of the hybrid protocol and discover a number of major design flaws which are fundamental to the design and seriously degrade the security of the system. In particular, we find that reuse of the cryptographic spreading sequence leads to a catastrophic failure wherein all satellites' data sequences can be recovered with high probability given knowledge of any single satellite's data sequence. This also enables sufficient recovery of the spreading sequence to spoof arbitrary messages, and increases vulnerability to optimized jamming. We evaluate and validate these findings through simulations with respect to real-world systems, and use this to propose countermeasures and system improvements which should be considered as standardization work continues.

As the use of satellites continues to grow, new networking paradigms are emerging to support the scale and long distance communication inherent to these networks. In particular, interplanetary communication relays connect distant network segments together, but result in a sparsely connected network with long-distance links that are frequently interrupted. In this new context, traditional Public Key Infrastructure (PKI) becomes difficult to implement, due to the impossibility of low-latency queries to a central authority. This paper addresses the challenge of implementing PKI in these complex networks, identifying the essential goals and requirements. Using these requirements, we develop the KeySpace framework, comprising a set of standardized experiments and metrics for comparing PKI systems across various network topologies, evaluating their performance and security. This enables the testing of different protocols and configurations in a standard, repeatable manner, so that improvements can be more fairly tested and clearly demonstrated. We use KeySpace to test two standard PKI protocols in use in terrestrial networks (OCSP and CRLs), demonstrating for the first time that both can be effectively utilized even in interplanetary networks with high latency and frequent interruptions, provided authority is properly distributed throughout the network. Finally, we propose and evaluate a number of novel techniques extending standard OCSP to improve the overhead of connection establishment, reduce link congestion, and limit the reach of an attacker with a compromised key. Using KeySpace we validate these claims, demonstrating their improved performance over the state of the art.

In the wake of increasing numbers of attacks on radio communication systems, a range of techniques are being deployed to increase the security of these systems. One such technique is radio fingerprinting, in which the transmitter can be identified and authenticated by observing small hardware differences expressed in the signal. Fingerprinting has been explored in particular in the defense of satellite systems, many of which are insecure and cannot be retrofitted with cryptographic security. In this paper, we evaluate the effectiveness of radio fingerprinting techniques under interference and jamming attacks, usually intended to deny service. By taking a pre-trained fingerprinting model and gathering a new dataset in which different levels of Gaussian noise and tone jamming have been added to the legitimate signal, we assess the attacker power required in order to disrupt the transmitter fingerprint such that it can no longer be recognized. We compare this to Gaussian jamming on the data portion of the signal, obtaining the remarkable result that transmitter fingerprints are still recognizable even in the presence of moderate levels of noise. Through deeper analysis of the results, we conclude that it takes a similar amount of jamming power in order to disrupt the fingerprint as it does to jam the message contents itself, so it is safe to include a fingerprinting system to authenticate satellite communication without opening up the system to easier denial-of-service attacks.

In recent years, numerous sophisticated malware detection systems have been proposed, many of which are based on machine learning. Though such systems attain impressive results, they are often designed having effectiveness as the main, if not only, requirement. As a result, the effectiveness of such systems, especially if based on deep learning models, often comes with (i) poor extensibility, being very difficult to adapt and/or extend to other settings, and (ii) poor explainability, since it is often not possible for humans to understand the reasons behind the model’s predictions, making further analysis of threats a challenge. In this paper we show how it is possible to design an extensible and explainable yet effective malware detection system. Extensibility is obtained thanks to the exploitation of TTPs (Tactics, Techniques, and Procedures) from the popular MITRE ATT&CK framework, which is an ontology of adversarial behaviour that allows us to divide the general problem of malware detection into the smaller problems of detecting the different types of malicious activity that can be carried out. Explainability is obtained by returning (i) which TTPs have been detected and are responsible for the classification of the entire behaviour as malicious, and (ii) why such TTPs have been classified as malicious. To demonstrate the viability of this approach we implement these ideas in a system called RADAR. We evaluate RADAR on a very large dataset comprising of 2,286,907 malicious and benign samples, representing a total of 84,792,452 network flows. The experimental analysis confirms that the proposed methodology can be effectively exploited: RADAR’s ability to detect malware is comparable to other state-of-the-art non-interpretable systems’ capabilities. To the best of our knowledge, RADAR is the first TTP-based system for malware detection that uses machine learning while being extensible and explainable.

Adaptive Coding and Modulation (ACM) space protocols are currently being implemented and standardised to maximise data throughput on satellite links in the presence of signal fading and radio interference. These systems use an uplink feedback channel, allowing the ground segment to influence the selection of modulation and coding parameters to suit the current channel conditions. However, no current academic work considers the security implications of physical-layer attackers targeting this uplink channel.In this paper, we introduce the uplink-assisted attacker which hijacks the currently unauthenticated ACM feedback mechanisms, using only cheaply available equipment, to select illsuited communication parameters and prevent the channel from responding to radio interference attacks on the downlink. Our results show the high impact of this attack class: an uplink-assisted noise jammer can cause a 50% frame error rate at 11.5 dB less average power than a noise jammer alone, and up to 16.9 dB if higher modulation and coding parameters are supported. Uplink-assisted spoofing and bandwidth restricting attackers are also shown to be more effective than their counterparts which attack the downlink alone.Unfortunately, these issues cannot be resolved by cryptographic authentication alone, especially where an attacker can pose as one of multiple terminals reporting channel quality. We therefore conclude with a discussion of countermeasures to prevent and detect this form of attack, and draw out lessons learned for secure ACM design.

Recent years have seen a rapid increase in the number of CubeSats and other small satellites in orbit - these have highly constrained computational and communication resources, but still require robust secure communication to operate effectively. The QUIC transport layer protocol is designed to provide efficient communication with cryptography guarantees built-in, with a particular focus on networks with high latency and packet loss. In this work we provide spaceQUIC, a proof of concept implementation of QUIC for NASA's "core Flight System" satellite operating system, and assess its performance.

In this paper, we show that by using inertial sensor data generated by a smart ring, worn on the finger, the user can be authenticated when making mobile payments or when knocking on a door (for access control). The proposed system can be deployed purely in software and does not require updates to existing payment terminals or infrastructure. We also demonstrate that smart ring data can authenticate smartwatch gestures, and vice versa, allowing either device to act as an implicit second factor for the other. To validate the system, we conduct a user study (n=21) to collect inertial sensor data from users as they perform gestures, and we evaluate the system against an active impersonation attacker. Based on this data, we develop payment and access control authentication models for which we achieve EERs of 0.04 and 0.02, respectively.

Electric vehicle charging sessions can be authorised in different ways, ranging from smartphone applications to smart cards with unique identifiers that link the electric vehicle to the charging station.However, these methods do not provide strong authentication guarantees.In this paper, we propose a novel second factor authentication scheme to tackle this problem.We show that by using inertial sensor data collected from IMU sensors either embedded in the handle of the charging cable or on a separate smartwatch, users can be authenticated implicitly by behavioural biometrics as they unhook the cable from the charging station and plug it into their car at the start of a charging session.To validate the system, we conducted a user study (n=20) to collect data and we developed a suite of authentication models for which we achieve EERs of 0.06.

Network analysis and machine learning techniques have been widely applied for building malware detection systems. Though these systems attain impressive results, they often are (i) not extensible, being monolithic, well tuned for the specific task they have been designed for but very difficult to adapt and/or extend to other settings, and (ii) not interpretable, being black boxes whose inner complexity makes it impossible to link the result of detection with its root cause, making further analysis of threats a challenge. In this paper we present RADAR, an extensible and explainable system that exploits the popular TTP (Tactics, Techniques, and Procedures) ontology of adversary behaviour described in the industry-standard MITRE ATT&CK framework in order to unequivocally identify and classify malicious behaviour using network traffic. We evaluate RADAR on a very large dataset comprising of 2,286,907 malicious and benign samples, representing a total of 84,792,452 network flows. The experimental analysis confirms that the proposed methodology can be effectively exploited: RADAR’s ability to detect malware is comparable to other state-of-the-art non-interpretable systems’ capabilities. To the best of our knowledge, RADAR is the first TTP-based system for malware detection that uses machine learning while being extensible and explainable.

In the last decade, machine learning (ML) methods have increasingly been applied to the task of malware detection. While these approaches have surely demonstrated their effectiveness, they still present limitations, some of which are a consequence of their purely data-driven nature. In this paper, we show how the MITRE ATT&CK framework of tactics, techniques, and procedures (TTPs) can be exploited to overcome such limitations and improve their ability to detect malware on networks. We conduct an extensive experimental analysis, testing 7 ML models on 5 large datasets comprising over 37 million flows. Our results clearly demonstrate that adding TTP-based features for training the models robustly improves their performance. Our models outperform the standard ones 922 times out of a total of 952, (i.e., 96.8% of the time), with the biggest improvements (up to 84.9% in terms of FPR) being observed in situations designed to be challenging for ML models.

Satellite user terminals are a promising target for adversaries seeking to target satellite communication networks. Despite this, many protections commonly found in terrestrial routers are not present in some user terminals. As a case study we audit the attack surface presented by the Starlink router's admin interface, using fuzzing to uncover a denial of service attack on the Starlink user terminal. We explore the attack's impact, particularly in the cases of drive-by attackers, and attackers that are able to maintain a continuous presence on the network. Finally, we discuss wider implications, looking at lessons learned in terrestrial router security, and how to properly implement them in this new context.

Despite their substantially different purposes, a common issue both for many legacy data links and many onboard data buses used in aviation is a lack of authentication and thus a vulnerability to spoofing and message manipulation attacks. This problem has been discussed at length for some prominent technologies of both sorts (e.g., ADS-B, ARINC 429) but is common in many more cases (e.g., ACARS, CPDLC, MIL-STD-1553, RS-485). For ground data links, attacks can take place over-the-air, while for onboard buses an attacker requires some physical access to an aircraft. Yet, in both cases an attacker’s goal is to obtain control of a transceiver on the communication channel. As such, hardware fingerprinting methods are useful in both contexts. Prior work has applied such methods to specific protocols. We now propose a transferable fingerprinting scheme that has applicability in both contexts, describe our experiences in applying it for each and evaluate its performance in benign and malicious conditions. As replacing legacy communication links with newer, more secure protocols is practically challenging, the improvements to practical deployment offered by our system represent a meaningful benefit for real deployment efforts.