Georg Carle

Tianlun Hu, Qi Liao, Qiang Liu et al.

Network slicing enables operators to efficiently support diverse applications on a common physical infrastructure. The ever-increasing densification of network deployment leads to complex and non-trivial inter-cell interference, which requires more than inaccurate analytic models to dynamically optimize resource management for network slices. In this paper, we develop a DIRP algorithm with multiple deep reinforcement learning (DRL) agents to cooperatively optimize resource partition in individual cells to fulfill the requirements of each slice, based on two alternative reward functions. Nevertheless, existing DRL approaches usually tie the pretrained model parameters to specific network environments with poor transferability, which raises practical deployment concerns in large-scale mobile networks. Hence, we design a novel transfer learning-aided DIRP (TL-DIRP) algorithm to ease the transfer of DIRP agents across different network environments in terms of sample efficiency, model reproducibility, and algorithm scalability. The TL-DIRP algorithm first centrally trains a generalized model and then transfers the "generalist" to each local agent as "specialist" with distributed finetuning and execution. TL-DIRP consists of two steps: 1) centralized training of a generalized distributed model, 2) transferring the "generalist" to each "specialist" with distributed finetuning and execution. The numerical results show that not only DIRP outperforms existing baseline approaches in terms of faster convergence and higher reward, but more importantly, TL-DIRP significantly improves the service performance, with reduced exploration cost, accelerated convergence rate, and enhanced model reproducibility. As compared to a traffic-aware baseline, TL-DIRP provides about 15% less violation ratio of the quality of service (QoS) for the worst slice service and 8.8% less violation on the average service QoS.

Tianlun Hu, Qi Liao, Qiang Liu et al.

Deep reinforcement learning (DRL) has been increasingly employed to handle the dynamic and complex resource management in network slicing. The deployment of DRL policies in real networks, however, is complicated by heterogeneous cell conditions. In this paper, we propose a novel transfer learning (TL) aided multi-agent deep reinforcement learning (MADRL) approach with inter-agent similarity analysis for inter-cell inter-slice resource partitioning. First, we design a coordinated MADRL method with information sharing to intelligently partition resource to slices and manage inter-cell interference. Second, we propose an integrated TL method to transfer the learned DRL policies among different local agents for accelerating the policy deployment. The method is composed of a new domain and task similarity measurement approach and a new knowledge transfer approach, which resolves the problem of from whom to transfer and how to transfer. We evaluated the proposed solution with extensive simulations in a system-level simulator and show that our approach outperforms the state-of-the-art solutions in terms of performance, convergence speed and sample efficiency. Moreover, by applying TL, we achieve an additional gain over 27% higher than the coordinate MADRL approach without TL.

Wenxuan Ye, Chendi Qian, Xueli An et al.

Integrating native AI support into the network architecture is an essential objective of 6G. Federated Learning (FL) emerges as a potential paradigm, facilitating decentralized AI model training across a diverse range of devices under the coordination of a central server. However, several challenges hinder its wide application in the 6G context, such as malicious attacks and privacy snooping on local model updates, and centralization pitfalls. This work proposes a trusted architecture for supporting FL, which utilizes Distributed Ledger Technology (DLT) and Graph Neural Network (GNN), including three key features. First, a pre-processing layer employing homomorphic encryption is incorporated to securely aggregate local models, preserving the privacy of individual models. Second, given the distributed nature and graph structure between clients and nodes in the pre-processing layer, GNN is leveraged to identify abnormal local models, enhancing system security. Third, DLT is utilized to decentralize the system by selecting one of the candidates to perform the central server's functions. Additionally, DLT ensures reliable data management by recording data exchanges in an immutable and transparent ledger. The feasibility of the novel architecture is validated through simulations, demonstrating improved performance in anomalous model detection and global model accuracy compared to relevant baselines.

Mostafa Abdelrahman, Filip Rezabek, Lars Hupel et al.

Digital signatures are crucial for securing Central Bank Digital Currencies (CBDCs) transactions. Like most forms of digital currencies, CBDC solutions rely on signatures for transaction authenticity and integrity, leading to major issues in the case of private key compromise. Our work explores threshold signature schemes (TSSs) in the context of CBDCs. TSSs allow distributed key management and signing, reducing the risk of a compromised key. We analyze CBDC-specific requirements, considering the applicability of TSSs, and use Filia CBDC solution as a base for a detailed evaluation. As most of the current solutions rely on ECDSA for compatibility, we focus on ECDSA-based TSSs and their supporting libraries. Our performance evaluation measured the computational and communication complexity across key processes, as well as the throughput and latency of end-to-end transactions. The results confirm that TSS can enhance the security of CBDC implementations while maintaining acceptable performance for real-world deployments.

Tianlun Hu, Qi Liao, Qiang Liu et al.

Network slicing is a key technique in 5G and beyond for efficiently supporting diverse services. Many network slicing solutions rely on deep learning to manage complex and high-dimensional resource allocation problems. However, deep learning models suffer limited generalization and adaptability to dynamic slicing configurations. In this paper, we propose a novel framework that integrates constrained optimization methods and deep learning models, resulting in strong generalization and superior approximation capability. Based on the proposed framework, we design a new neural-assisted algorithm to allocate radio resources to slices to maximize the network utility under inter-slice resource constraints. The algorithm exhibits high scalability, accommodating varying numbers of slices and slice configurations with ease. We implement the proposed solution in a system-level network simulator and evaluate its performance extensively by comparing it to state-of-the-art solutions including deep reinforcement learning approaches. The numerical results show that our solution obtains near-optimal quality-of-service satisfaction and promising generalization performance under different network slicing scenarios.

Wenxuan Ye, Xueli An, Onur Ayan et al.

Large models, renowned for superior performance, outperform smaller ones even without billion-parameter scales. While mobile network servers have ample computational resources to support larger models than client devices, privacy constraints prevent clients from directly sharing their raw data. Federated Learning (FL) enables decentralized clients to collaboratively train a shared model by exchanging model parameters instead of transmitting raw data. Yet, it requires a uniform model architecture and multiple communication rounds, which neglect resource heterogeneity, impose heavy computational demands on clients, and increase communication overhead. To address these challenges, we propose FedOL, to construct a larger and more comprehensive server model in one-shot settings (i.e., in a single communication round). Instead of model parameter sharing, FedOL employs knowledge distillation, where clients only exchange model prediction outputs on an unlabeled public dataset. This reduces communication overhead by transmitting compact predictions instead of full model weights and enables model customization by allowing heterogeneous model architectures. A key challenge in this setting is that client predictions may be biased due to skewed local data distributions, and the lack of ground-truth labels in the public dataset further complicates reliable learning. To mitigate these issues, FedOL introduces a specialized objective function that iteratively refines pseudo-labels and the server model, improving learning reliability. To complement this, FedOL incorporates a tailored pseudo-label generation and knowledge distillation strategy that effectively integrates diverse knowledge. Simulation results show that FedOL significantly outperforms existing baselines, offering a cost-effective solution for mobile networks where clients possess valuable private data but limited computational resources.

Sayantini Majumdar, Riccardo Trivisonno, Georg Carle

The management of networks is automated by closed loops. Concurrent closed loops aiming for individual optimization cause conflicts which, left unresolved, leads to significant degradation in performance indicators, resulting in sub-optimal network performance. Centralized optimization avoids conflicts, but impractical in large-scale networks for time-critical applications. Distributed, pervasive intelligence is therefore envisaged in the evolution to B5G networks. In this letter, we propose a Q-Learning-based distributed architecture (QLC), addressing the conflict issue by encouraging cooperation among intelligent agents. We design a realistic B5G network slice auto-scaling model and validate the performance of QLC via simulations, justifying further research in this direction.

Marton Kajo, Janik Schnellbach, Stephen S. Mwanje et al.

Deep learning will play a crucial role in enabling cognitive automation for the mobile networks of the future. Deep clustering, a subset of deep learning, could be a valuable tool for many network automation use-cases. Unfortunately, most state-of-the-art clustering algorithms target image datasets, which makes them hard to apply to mobile network data due to their highly tuned nature and related assumptions about the data. In this paper, we propose a new algorithm, DANCE (Decorrelating Adversarial Nets for Clustering-friendly Encoding), intended to be a reliable deep clustering method which also performs well when applied to network automation use-cases. DANCE uses a reconstructive clustering approach, separating clustering-relevant from clustering-irrelevant features in a latent representation. This separation removes unnecessary information from the clustering, increasing consistency and peak performance. We comprehensively evaluate DANCE and other select state-of-the-art deep clustering algorithms, and show that DANCE outperforms these algorithms by a significant margin on a mobile network dataset.

Marton Kajo, Stephen S. Mwanje, Benedek Schultz et al.

Deep Learning methods have been adopted in mobile networks, especially for network management automation where they provide means for advanced machine cognition. Deep learning methods utilize cutting-edge hardware and software tools, allowing complex cognitive algorithms to be developed. In a recent paper, we introduced the Bounding Sphere Quantization (BSQ) algorithm, a modification of the k-Means algorithm, that was shown to create better quantizations for certain network management use-cases, such as anomaly detection. However, BSQ required a significantly longer time to train than k-Means, a challenge which can be overcome with a neural network-based implementation. In this paper, we present such an implementation of BSQ that utilizes state-of-the-art deep learning tools to achieve a competitive training speed.

Holger Kinkelin, Richard von Seck, Christoph Rudolf et al.

The security of cryptographic communication protocols that use X.509 certificates depends on the correctness of those certificates. This paper proposes a system that helps to ensure the correct operation of an X.509 certification authority and its registration authorities. We achieve this goal by enforcing a policy-defined, multi-party validation and authorization workflow of certificate signing requests. Besides, our system offers full accountability for this workflow for forensic purposes. As a foundation for our implementation, we leverage the distributed ledger and smart contract framework Hyperledger Fabric. Our implementation inherits the strong tamper-resistance of Fabric which strengthens the integrity of the computer processes that enforce the validation and authorization of the certificate signing request, and of the metadata collected during certificate issuance.

Dominik Scholz, Sebastian Gallenmüller, Henning Stubbe et al.

The SYN flood attack is a common attack strategy on the Internet, which tries to overload services with requests leading to a Denial-of-Service (DoS). Highly asymmetric costs for connection setup - putting the main burden on the attackee - make SYN flooding an efficient and popular DoS attack strategy. Abusing the widely used TCP as an attack vector complicates the detection of malicious traffic and its prevention utilizing naive connection blocking strategies. Modern programmable data plane devices are capable of handling traffic in the 10 Gbit/s range without overloading. We discuss how we can harness their performance to defend entire networks against SYN flood attacks. Therefore, we analyze different defense strategies, SYN authentication and SYN cookie, and discuss implementation difficulties when ported to different target data planes: software, network processors, and FPGAs. We provide prototype implementations and performance figures for all three platforms. Further, we fully disclose the artifacts leading to the experiments described in this work.

Michel Barbeau, Georg Carle, Joaquin Garcia-Alfaro et al.

Cyber-Physical Systems (CPS) consist of distributed engineered environments where the monitoring and surveillance tasks are governed by tightly integrated computing, communication and control technologies. CPS are omnipresent in our everyday life. Hacking and failures of such systems have impact on critical services with potentially significant and lasting consequences. In this paper, we review which requirements a CPS must meet to address the challenges of tomorrow. Two key challenges are understanding and reinforcing the resilience of CPS.

Sebastian Gallenmüller, Dominik Schöffmann, Dominik Scholz et al.

Secure communication is an integral feature of many Internet services. The widely deployed TLS protects reliable transport protocols. DTLS extends TLS security services to protocols relying on plain UDP packet transport, such as VoIP or IoT applications. In this paper, we construct a model to determine the performance of generic DTLS-enabled applications. Our model considers basic network characteristics, e.g., number of connections, and the chosen security parameters, e.g., the encryption algorithm in use. Measurements are presented demonstrating the applicability of our model. These experiments are performed using a high-performance DTLS-enabled VPN gateway built on top of the well-established libraries DPDK and OpenSSL. This VPN solution represents the most essential parts of DTLS, creating a DTLS performance baseline. Using this baseline the model can be extended to predict even more complex DTLS protocols besides the measured VPN. Code and measured data used in this paper are publicly available at https://git.io/MoonSec and https://git.io/Sdata.

Holger Kinkelin, Heiko Niedermayer, Marc Müller et al.

Configuration management in networks with highest security demands must not depend on just one administrator and her device. Otherwise, problems can be caused by mistakes or malicious behavior of this admin, or when her computer got compromised, which allows an attacker to abuse the administrator's far-reaching permissions. Instead, we propose to use a reliable and resilient configuration management process orchestrated by a configuration management system (CMS). This can be achieved by separation of concerns (proposing a configuration vs. authorizing it), employing multi-party authorization (MPA), and enforcing that only authorized configurations can be deployed. This results in a configuration management process that is decentralized on a human, decision-making level, and a technical, device level. However, due to different opinions or adversarial interference, the result of an MPA process can end in a conflict. This raises the question how such conflicts can be mediated in a better way than just employing majority voting, which is insufficient in certain situations. As an alternative, this paper introduces building blocks of customizable conflict mediation strategies which we integrated into our CMS TANCS . The conflict mediation functionality as well as the initial TANCS implementation run on top of the distributed ledger and smart contract framework Hyperledger Fabric which makes all processes resilient and tamper-resistant.

Marcel von Maltitz, Dominik Bitzer, Georg Carle

In the Internet of Things and smart environments data, collected from distributed sensors, is typically stored and processed by a central middleware. This allows applications to query the data they need for providing further services. However, centralization of data causes several privacy threats: The middleware becomes a third party which has to be trusted, linkage and correlation of data from different context becomes possible and data subject lose control over their data. Hence, other approaches than centralized processing should be considered. Here, Secure Multiparty Computation is a promising candidate for secure and privacy-preserving computation happening close to the sources of the data. In order to make SMC fit for application in these contexts, we extend SMC to act as a service: We provide elements which allow third parties to query computed data from a group of peers performing SMC. Furthermore, we establish fine-granular access control on the level of individual data queries, yielding data protection of the computed results. By adding measures to inform data sources about requests and the usage of their data, we show how a fully privacy-preserving service can be built on the foundation of SMC.

Quirin Scheitle, Oliver Gasser, Theodor Nolte et al.

In this paper, we analyze the evolution of Certificate Transparency (CT) over time and explore the implications of exposing certificate DNS names from the perspective of security and privacy. We find that certificates in CT logs have seen exponential growth. Website support for CT has also constantly increased, with now 33% of established connections supporting CT. With the increasing deployment of CT, there are also concerns of information leakage due to all certificates being visible in CT logs. To understand this threat, we introduce a CT honeypot and show that data from CT logs is being used to identify targets for scanning campaigns only minutes after certificate issuance. We present and evaluate a methodology to learn and validate new subdomains from the vast number of domains extracted from CT logged certificates.

Marcel von Maltitz, Georg Carle

Centralized systems in the Internet of Things---be it local middleware or cloud-based services---fail to fundamentally address privacy of the collected data. We propose an architecture featuring secure multiparty computation at its core in order to realize data processing systems which already incorporate support for privacy protection in the architecture.

Holger Kinkelin, Valentin Hauner, Heiko Niedermayer et al.

Numerous IoT applications, like building automation or process control of industrial sites, exist today. These applications inherently have a strong connection to the physical world. Hence, IT security threats cannot only cause problems like data leaks but also safety issues which might harm people. Attacks on IT systems are not only performed by outside attackers but also insiders like administrators. For this reason, we present ongoing work on a configuration management system (CMS) that provides control over administrators, restrains their rights, and enforces separation of concerns. We reach this goal by conducting a configuration management process that requires multi-party authorization for critical configurations to achieve Byzantine fault tolerance against attacks and faults by administrators. Only after a configuration has been authorized by multiple experts, it is applied to the targeted devices. For the whole configuration management process, our CMS guarantees accountability and traceability. Lastly, our system is tamper-resistant as we leverage Hyperledger Fabric, which provides a distributed execution environment for our CMS and a blockchain-based distributed ledger that we use to store the configurations. A beneficial side effect of this approach is that our CMS is also suitable to manage configurations for infrastructure shared across different organizations that do not need to trust each other.

Benjamin Hof, Georg Carle

A large user base relies on software updates provided through package managers. This provides a unique lever for improving the security of the software update process. We propose a transparency system for software updates and implement it for a widely deployed Linux package manager, namely APT. Our system is capable of detecting targeted backdoors without producing overhead for maintainers. In addition, in our system, the availability of source code is ensured, the binding between source and binary code is verified using reproducible builds, and the maintainer responsible for distributing a specific package can be identified. We describe a novel "hidden version" attack against current software transparency systems and propose as well as integrate a suitable defense. To address equivocation attacks by the transparency log server, we introduce tree root cross logging, where the log's Merkle tree root is submitted into a separately operated log server. This significantly relaxes the inter-operator cooperation requirements compared to other systems. Our implementation is evaluated by replaying over 3000 updates of the Debian operating system over the course of two years, demonstrating its viability and identifying numerous irregularities.

Marcel von Maltitz, Cornelius Diekmann, Georg Carle

Privacy analysis is critical but also a time-consuming and tedious task. We present a formalization which eases designing and auditing high-level privacy properties of software architectures. It is incorporated into a larger policy analysis and verification framework and enables the assessment of commonly accepted data protection goals of privacy. The formalization is based on static taint analysis and makes flow and processing of privacy-critical data explicit, globally as well as on the level of individual data subjects. Formally, we show equivalence to traditional label-based information flow security and prove overall soundness of our tool with Isabelle/HOL. We demonstrate applicability in two real-world case studies, thereby uncovering previously unknown violations of privacy constraints in the analyzed software architectures.

Johann Schlamp, Ralph Holz, Quentin Jacquemart et al.

The detection of BGP prefix hijacking attacks has been the focus of research for more than a decade. However, state-of-the-art techniques fall short of detecting more elaborate types of attack. To study such attacks, we devise a novel formalization of Internet routing, and apply this model to routing anomalies in order to establish a comprehensive attacker model. We use this model to precisely classify attacks and to evaluate their impact and detectability. We analyze the eligibility of attack tactics that suit an attacker's goals and demonstrate that related work mostly focuses on less impactful kinds of attacks. We further propose, implement and test the Hijacking Event Analysis Program (HEAP), a new approach to investigate hijacking alarms. Our approachis designed to seamlessly integrate with previous work in order to reduce the high rates of false alarms inherent to these techniques. We leverage several unique data sources that can reliably disprove malicious intent. First, we make use of an Internet Routing Registry to derive business or organisational relationships between the parties involved in an event. Second, we use a topology-based reasoning algorithm to rule out events caused by legitimate operational practice. Finally, we use Internet-wide network scans to identify SSL/TLS-enabled hosts, which helps to identify non-malicious events by comparing public keys prior to and during an event. In our evaluation, we prove the effectiveness of our approach, and show that day-to-day routing anomalies are harmless for the most part. More importantly, we use HEAP to assess the validity of publicly reported alarms. We invite researchers to interface with HEAP in order to cross-check and narrow down their hijacking alerts.

Johann Schlamp, Matthias Wählisch, Thomas C. Schmidt et al.

The Internet routing protocol BGP expresses topological reachability and policy-based decisions simultaneously in path vectors. A complete view on the Internet backbone routing is given by the collection of all valid routes, which is infeasible to obtain due to information hiding of BGP, the lack of omnipresent collection points, and data complexity. Commonly, graph-based data models are used to represent the Internet topology from a given set of BGP routing tables but fall short of explaining policy contexts. As a consequence, routing anomalies such as route leaks and interception attacks cannot be explained with graphs. In this paper, we use formal languages to represent the global routing system in a rigorous model. Our CAIR framework translates BGP announcements into a finite route language that allows for the incremental construction of minimal route automata. CAIR preserves route diversity, is highly efficient, and well-suited to monitor BGP path changes in real-time. We formally derive implementable search patterns for route leaks and interception attacks. In contrast to the state-of-the-art, we can detect these incidents. In practical experiments, we analyze public BGP data over the last seven years.

Cornelius Diekmann, Lars Hupel, Georg Carle

Large systems are commonly internetworked. A security policy describes the communication relationship between the networked entities. The security policy defines rules, for example that A can connect to B, which results in a directed graph. However, this policy is often implemented in the network, for example by firewalls, such that A can establish a connection to B and all packets belonging to established connections are allowed. This stateful implementation is usually required for the network's functionality, but it introduces the backflow from B to A, which might contradict the security policy. We derive compliance criteria for a policy and its stateful implementation. In particular, we provide a criterion to verify the lack of side effects in linear time. Algorithms to automatically construct a stateful implementation of security policy rules are presented, which narrows the gap between formalization and real-world implementation. The solution scales to large networks, which is confirmed by a large real-world case study. Its correctness is guaranteed by the Isabelle/HOL theorem prover.