TY - THES A1 - Danner, Dominik T1 - Towards Quality of Service and Fairness in Smart Grid Applications N2 - Due to the increasing amount of distributed renewable energy generation and the emerging high demand at consumer connection points, e. g., electric vehicles, the power distribution grid will reach its capacity limit at peak load times if it is not expensively enhanced. Alternatively, smart flexibility management that controls user assets can help to better utilize the existing power grid infrastructure for example by sharing available grid capacity among connected electric vehicles or by disaggregating flexibility requests to hybrid photovoltaic battery energy storage systems in households. Besides maintaining an acceptable state of the power distribution grid, these smart grid applications also need to ensure a certain quality of service and provide fairness between the individual participants, both of which are not extensively discussed in the literature. This thesis investigates two smart grid applications, namely electric vehicle charging-as-a-service and flexibility-provision-as-a-service from distributed energy storage systems in private households. The electric vehicle charging service allocation is modeled with distributed queuing-based allocation mechanisms which are compared to new probabilistic algorithms. Both integrate user constraints (arrival time, departure time, and energy required) to manage the quality of service and fairness. In the queuing-based allocation mechanisms, electric vehicle charging requests are packetized into logical charging current packets, representing the smallest controllable size of the charging process. These packets are queued at hierarchically distributed schedulers, which allocate the available charging capacity using the time and frequency division multiplexing technique known from the networking domain. This allows multiple electric vehicles to be charged simultaneously with variable charging currents. To achieve high quality of service and fairness among electric vehicle charging processes, dynamic weights are introduced into a weighted fair queuing scheduler that considers electric vehicle departure time and required energy for prioritization. The distributed probabilistic algorithms are inspired by medium access protocols from computer networking, such as binary exponential backoff, and control the quality of service and fairness by adjusting sampling windows and waiting periods based on user requirements. The second smart grid application under investigation aims to provide flexibility provision-as-a-service that disaggregates power flexibility requests to distributed battery energy storage systems in private households. Commonly, the main purpose of stationary energy storage is to store energy from a local photovoltaic system for later use, e. g., for overnight charging of an electric vehicle. This is optimized locally by a home energy management system, which also allows the scheduling of external flexibility requests defined by the deviation from the optimal power profile at the grid connection point, for example, to perform peak shaving at the transformer. This thesis discusses a linear heuristic and a meta heuristic to disaggregate a flexibility request to the single participating energy management systems that are grouped into a flexibility pool. Thereby, the linear heuristic iteratively assigns portions of the power flexibility to the most appropriate energy management system for one time slot after another, minimizing the total flexibility cost or maximizing the probability of flexibility delivery. In addition, a multi-objective genetic algorithm is proposed that also takes into account power grid aspects, quality of service, and fairness among par-ticipating households. The genetic operators are tailored to the flexibility disaggregation search space, taking into account flexibility and energy management system constraints, and enable power-optimized buffering of fitness values. Both smart grid applications are validated on a realistic power distribution grid with real driving patterns and energy profiles for photovoltaic generation and household consumption. The results of all proposed algorithms are analyzed with respect to a set of newly defined metrics on quality of service, fairness, efficiency, and utilization of the power distribution grid. One of the main findings is that none of the tested algorithms outperforms the others in all quality of service metrics, however, integration of user expectations improves the service quality compared to simpler approaches. Furthermore, smart grid control that incorporates users and their flexibility allows the integration of high-load applications such as electric vehicle charging and flexibility aggregation from distributed energy storage systems into the existing electricity distribution infrastructure. However, there is a trade-off between power grid aspects, e. g., grid losses and voltage values, and the quality of service provided. Whenever active user interaction is required, means of controlling the quality of service of users’ smart grid applications are necessary to ensure user satisfaction with the services provided. Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-13731 ER - TY - THES A1 - Brummer, Stephan T1 - Numerisch robuste Berechnung der zirkulären Sichtbarkeitsmenge N2 - Sichtbarkeitsprobleme, wie das Folgende, gehören zu den grundlegenden Problemen der algorithmischen Geometrie: Berechne zu einem einfachen Polygon, dem sogenannten Kanal, und zu einem darin enthaltenen Punkt die von diesem Punkt aus sichtbare Punktmenge. Dabei ist ein Punkt von einem anderen Punkt aus sichtbar, wenn deren Verbindungsstrecke den Kanal nicht verlässt. Wir wollen uns in dieser Arbeit mit zirkulärer Sichtbarkeit beschäftigen. Zur Verbindung zweier Punkte sind dann nicht nur Strecken, sondern auch Kreisbögen zulässig. Außerdem betrachten wir als Ausgangspunkt dieser sogenannten Sichtbarkeitskreisbögen und -strecken eine Kante des Kanals anstatt eines einzelnen Punkts. Konkret liefert diese Arbeit einen Beitrag zur numerisch robusten Bestimmung der zirkulären Sichtbarkeitsmenge ausgehend von einer Kante des Kanals. Hierfür wird in dieser Arbeit ein Algorithmus vorgestellt, mit dem für einen gegebenen Punkt festgestellt werden kann, ob dieser von der Startkante aus sichtbar ist. Im Fall eines sichtbaren Punkts wird ein Sichtbarkeitskreisbogen berechnet, der zwei Kanalberührungen besitzt. Damit kann der Algorithmus bei geeigneter Wahl des zu untersuchenden Punkts – der als dritte Kanalberührung fungiert – direkt zur Berechnung von sogenannten Grenzkreisbögen der Sichtbarkeitsmenge benutzt werden. Diese definieren den Rand der zirkulären Sichtbarkeitsmenge und zeichnen sich dadurch aus, dass sie vom Kanal dreimal abwechselnd von links und von rechts berührt werden. Der beschriebene Algorithmus basiert auf der Untersuchung derjenigen Kreisbögen, die zwar nicht notwendigerweise vollständig im Kanal liegen, aber die Startkante mit dem Punkt verbinden, dessen Sichtbarkeit bestimmt werden soll. Insbesondere werden dabei die Bereiche untersucht, in denen der jeweilige Kreisbogen den Kanal verlässt, die sogenannten Verletzungen. Da die „Schwere“ einer solchen Verletzung quantifizierbar ist, wird ein iteratives Vorgehen ermöglicht. Dabei wird der Kreisbogen iterativ so verändert, dass dieser bei gleichem Endpunkt den Kanal immer „weniger verlässt“. Ist der Endpunkt und damit der zu untersuchende Punkt nicht sichtbar, wird im Laufe des Algorithmus festgestellt, dass keine derartige Verbesserung möglich ist. Der vorgestellte Algorithmus ist numerisch robust, einfach umzusetzen und besitzt eine in der Anzahl der Kanalecken lineare Laufzeit. Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-12299 ER - TY - JOUR A1 - Herbold, Steffen A1 - Hautli‑Janisz, Annette A1 - Heuer, Ute A1 - Kikteva, Zlata A1 - Trautsch, Alexander T1 - A large‑scale comparison of human‑written versus ChatGPT‑generated essays JF - Scientific Reports N2 - ChatGPT and similar generative AI models have attracted hundreds of millions of users and have become part of the public discourse. Many believe that such models will disrupt society and lead to significant changes in the education system and information generation. So far, this belief is based on either colloquial evidence or benchmarks from the owners of the models—both lack scientific rigor. We systematically assess the quality of AI-generated content through a large-scale study comparing human-written versus ChatGPT-generated argumentative student essays. We use essays that were rated by a large number of human experts (teachers). We augment the analysis by considering a set of linguistic characteristics of the generated essays. Our results demonstrate that ChatGPT generates essays that are rated higher regarding quality than human-written essays. The writing style of the AI models exhibits linguistic characteristics that are different from those of the human-written essays. Since the technology is readily available, we believe that educators must act immediately. We must re-invent homework and develop teaching concepts that utilize these AI models in the same way as math utilizes the calculator: teach the general concepts first and then use AI tools to free up time for other learning objectives. Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-13961 VL - 13 PB - Springer Nature ER - TY - JOUR A1 - Hassen, Wiem Fekih A1 - Ben Ahmed, Mariem T1 - Optimization of a Redox-Flow Battery Simulation Model Based on a Deep Reinforcement Learning Approach JF - Batteries N2 - Vanadium redox-flow batteries (VRFBs) have played a significant role in hybrid energy storage systems (HESSs) over the last few decades owing to their unique characteristics and advantages. Hence, the accurate estimation of the VRFB model holds significant importance in large-scale storage applications, as they are indispensable for incorporating the distinctive features of energy storage systems and control algorithms within embedded energy architectures. In this work, we propose a novel approach that combines model-based and data-driven techniques to predict battery state variables, i.e., the state of charge (SoC), voltage, and current. Our proposal leverages enhanced deep reinforcement learning techniques, specifically deep q-learning (DQN), by combining q-learning with neural networks to optimize the VRFB-specific parameters, ensuring a robust fit between the real and simulated data. Our proposed method outperforms the existing approach in voltage prediction. Subsequently, we enhance the proposed approach by incorporating a second deep RL algorithm—dueling DQN—which is an improvement of DQN, resulting in a 10% improvement in the results, especially in terms of voltage prediction. The proposed approach results in an accurate VFRB model that can be generalized to several types of redox-flow batteries. KW - energy storage KW - redox-flow battery KW - battery modeling KW - battery state variables KW - parameter optimization KW - accurate estimation KW - voltage prediction KW - deep reinforcement learning KW - deep q-learning KW - dueling deep q-networks Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-13994 VL - 10 PB - MDPI CY - Basel ER - TY - JOUR A1 - Hassen, Wiem Fekih A1 - Imen Azzouz, Imen Azzouz T1 - Optimization of Electric Vehicles Charging Scheduling Based on Deep Reinforcement Learning: A Decentralized Approach JF - Energies N2 - The worldwide adoption of Electric Vehicles (EVs) has embraced promising advancements toward a sustainable transportation system. However, the effective charging scheduling of EVs is not a trivial task due to the increase in the load demand in the Charging Stations (CSs) and the fluctuation of electricity prices. Moreover, other issues that raise concern among EV drivers are the long waiting time and the inability to charge the battery to the desired State of Charge (SOC). In order to alleviate the range of anxiety of users, we perform a Deep Reinforcement Learning (DRL) approach that provides the optimal charging time slots for EV based on the Photovoltaic power prices, the current EV SOC, the charging connector type, and the history of load demand profiles collected in different locations. Our implemented approach maximizes the EV profit while giving a margin of liberty to the EV drivers to select the preferred CS and the best charging time (i.e., morning, afternoon, evening, or night). The results analysis proves the effectiveness of the DRL model in minimizing the charging costs of the EV up to 60%, providing a full charging experience to the EV with a lower waiting time of less than or equal to 30 min. KW - smart EV charging KW - day-ahead planning KW - deep Q-Network KW - data-driven approach KW - waiting time KW - cost minimization KW - real dataset Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-13985 VL - 16 PB - MDPI CY - Basel ER - TY - JOUR A1 - Patil, Amit A1 - Ghasemi, Abdorasoul A1 - de Meer, Hermann T1 - Analysis of protection blinding in active distribution grids JF - IET Renewable Power Generation N2 - Protection blinding is a challenging issue in renewables-penetrated distribution grids and refers to a situation where a circuit breaker may not trip due to fault current contribution from distributed generation. This research addresses how the distributed generation location and capacity impact the operation of the circuit breaker in terms of the response time of the circuit breakers. The relative electrical distances of the faults and distributed generation to the circuit breakers are considered. The impact of distributed generation capacity considering the fault location is characterized using a new index called the heterogeneity index. The electrical distance between distributed generations and circuit breakers and the electrical distance between fault and circuit breaker is considered by a second new index called the electrical distance ratio. Data analysis on simulation results shows that these indices capture the phenomena of protection blinding caused by distributed generation. Results show that a higher distributed generation penetration and faults that are electri cally further away from a circuit breaker show severe cases of protection blinding captured by the indices. Furthermore, it is demonstrated how these indices can identify the worst impacted locations in the distribution grid. A key result is that protection blinding does not necessarily occur solely due to the presence of distributed generation between a circuit breaker and a fault, but is dependent on factors such as distributed generation location in the distribution grid, fault level, fault level distribution across the generation units and fault location. KW - General & introductory electrical electronics engineering Y1 - 2024 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-14661 ER - TY - THES A1 - Püllen, Dominik T1 - Holistic Security Engineering for Software-Defined Vehicles N2 - With the increasing use of digital technologies in the automotive sector, the traditional automobile is undergoing a structural transformation, requiring new technologies and enabling innovative mobility concepts. In particular, the ability to drive automatically or even fully autonomously, update control software, and remain connected to the environment allows attackers to infiltrate highly critical vehicle systems and take control without adequate protection. Once not only individual vehicles but entire fleets are dominated by software, cyberattacks could disrupt a significant portion of the infrastructure and expose passengers to substantial risks. This work follows a holistic approach to protecting highly automated software-defined vehicles from cyberattacks by designing and implementing security concepts in the main phases of a vehicle's lifecycle. We use SAE level 4 prototype vehicles to evaluate our proposed techniques. We start with a systematic security requirement analysis using the ISA-62443 standard series, demonstrating how threats can be identified in a collaborative, hierarchical process and how the resulting security risks impact the software and hardware architecture of a self-driving vehicle. We show how this analysis process results in concrete requirements whose consideration reduces the overall security risk to a tolerable level. Subsequently, we develop technical solutions for selected requirements. We begin by securing the CAN and FlexRay legacy protocols, which we foresee being used in specific areas of SDV in a transitional period despite technological changes. To enable vehicle-wide security management, we address the management and distribution of cryptographic keys within such networks, mainly focusing on resource-constrained devices. We propose using lightweight implicit certificates for deriving cryptographic group keys that can be used in CAN networks. Additionally, we demonstrate how the slot-based frame structure of the FlexRay protocol allows for efficient "multi-slot" authentication, for which we calculate cryptographic keys using hash-based key chains. SDV use Ethernet-based communication protocols and custom middleware stacks to transmit large amounts of data in real-time. We develop a three-stage security process for the novel ASOA, which enables the development and central orchestration of system-agnostic functional software components on embedded systems and HPC platforms. After the central specification of the security architecture at the data flow level, security tokens are automatically calculated and distributed for runtime protection of the service-oriented, DDS-based data transmission. Our process ensures the strict separation of function and system knowledge, allowing for cost-effective and adaptable security architecture management. The evaluation in four self-driving, software-defined vehicles demonstrates an average runtime overhead of approximately 5.71%. As the initial risk analysis and actual cyberattacks have shown, protective measures against the compromise of control units must be taken alongside communication security. To address this, we develop a method for verifying and validating the software integrity of control units. A governmental third party confirms a measurement through a digital certificate, proving the examined vehicle's trustworthiness and suitability for participation in automated traffic. In the final step of this work, we present an assessment scheme that allows software-defined vehicles to evaluate security incidents during operation in terms of their maximum expected damage and initiate appropriate countermeasures. We follow the ISO/SAE 21434 standard and model attack paths using a graph representing dependencies among internal vehicle assets to account for the propagation effects of cyberattacks. The assessment of a security incident considers not only the probability of individual attack paths but also the vehicle context. Our practical evaluation demonstrates that we can detect, report, and assess security incidents below the human reaction time in the earlier mentioned prototype vehicles. Y1 - 2024 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-14497 ER - TY - THES A1 - Alhamzeh, Alaa T1 - Language Reasoning by means of Argument Mining and Argument Quality N2 - Understanding of financial data has always been a point of interest for market participants to make better informed decisions. Recently, different cutting edge technologies have been addressed in the Financial Technology (FinTech) domain, including numeracy understanding, opinion mining and financial ocument processing. In this thesis, we are interested in analyzing the arguments of financial experts with the goal of supporting investment decisions. Although various business studies confirm the crucial role of argumentation in financial communications, no work has addressed this problem as a computational argumentation task. In other words, the automatic analysis of arguments. In this regard, this thesis presents contributions in the three essential axes of theory, data, and evaluation to fill the gap between argument mining and financial text. First, we propose a method for determining the structure of the arguments stated by company representatives during the public announcement of their quarterly results and future estimations through earnings conference calls. The proposed scheme is derived from argumentation theory at the micro-structure level of discourse. We further conducted the corresponding annotation study and published the first financial dataset annotated with arguments: FinArg. Moreover, we investigate the question of evaluating the quality of arguments in this financial genre of text. To tackle this challenge, we suggest using two levels of quality metrics, considering both the Natural Language Processing (NLP) literature of argument quality assessment and the financial era peculiarities. Hence, we have also enriched the FinArg data with our quality dimensions to produce the FinArgQuality dataset. In terms of evaluation, we validate the principle of ensemble learning on the argument identification and argument unit classification tasks. We show that combining a traditional machine learning model along with a deep learning one, via an integration model (stacking), improves the overall performance, especially in small dataset settings. In addition, despite the fact that argument mining is mainly a domain dependent task, to this date, the number of studies that tackle the generalization of argument mining models is still relatively small. Therefore, using our stacking approach and in comparison to the transfer learning model of DistilBert, we address and analyze three real-world scenarios concerning the model robustness over completely unseen domains and unseen topics. Furthermore, with the aim of the automatic assessment of argument strength, we have investigated and compared different (refined) versions of Bert-based models that incorporate external knowledge in the decision layer. Consequently, our method outperforms the baseline model by 13 ± 2% in terms of F1-score through integrating Bert with encoded categorical features. Beyond our theoretical and methodological proposals, our model of argument quality assessment, annotated corpora, and evaluation approaches are publicly available, and can serve as strong baselines for future work in both FinNLP and computational argumentation domains. Hence, directly exploiting this thesis, we proposed to the community, a new task/challenge related to the analysis of financial arguments: FinArg-1, within the framework of the NTCIR-17 conference. We also used our proposals to react to the Touché challenge at the CLEF 2021 conference. Our contribution was selected among the «Best of Labs». KW - NLP, Argument Mining, Argument Quality Assessment, Financial Argumentation, Earnings Conference Calls Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-12699 ER - TY - THES A1 - Graßl, Isabella T1 - Diversity in Programming Education: Effects of Topic and Group Constellation on Young Programming Novices N2 - The field of software engineering faces a significant diversity crisis, characterized by a critical lack of heterogeneity despite ongoing efforts to promote gender equality. The persistent male dominance in this domain has created an urgent need for more heterogeneous groups in software engineering. This lack of diversity not only hinders underrepresented groups from entering the field but also prevents them from gaining initial programming experiences, which are a core component of software engineering and essential for developing computational thinking. To address this crisis and its implications, early interventions are key in shaping positive perceptions, building confidence, and sparking initial interest in programming among underrepresented groups before societal stereotypes of programming as a nerdy field manifests. This means starting with basic programming courses for children and continuing through to first-year university students in order to foster technical skills and computational thinking, alongside creativity and collaboration. However, there is limited understanding of how introductory programming course designs impact diversity-dependent characteristics to create welcoming and learning-friendly environments. This understanding is particularly important for underrepresented groups, especially girls, to benefit from their first programming experiences as they are often hindered by the initial perception of programming as (1) abstract and unappealing, and (2) non-social to novices. Engaging, creative, and relatable topics in programming courses might demystify complex programming concepts, making them more accessible, less intimidating, and appealing. However, understanding programming is not just about the content---it is also about the context in which it is learned. Introducing programming as social activity is important, particularly for young learners. By emphasizing team work, we might encourage collaboration and peer support, counteracting the lone-wolf programmer stereotype. Therefore, this doctoral thesis investigates the effects of both key aspects in programming courses---(1) topic choices and (2) group constellations---on young programming novices. The aim is to provide a holistic understanding of how different course designs can support diverse learners and promote gender equality in programming education. While this research primarily addresses gender diversity due to the persistent gender gap in software engineering, it also examines additional diversity dimensions, including age, ethnicity, prior programming experience, disabilities, and educational background. A total of 13 studies were conducted within this thesis, examining the current state of educational settings and utilizing various introductory programming courses designed for children aged 8 to 18, as well as first-year university students. These studies employed different programming environments, such as Scratch and Sonic Pi, and incorporated a variety of topics and group constellations to observe their effects on student outcomes. By using a mixed-methods design, data were gathered through surveys, observations, and both data-driven and manual code analysis. Key findings reveal that it is particularly noteworthy how children utilize the programming environment to engage with and creatively express topics aligned with their interests which also align mostly with gender-stereotypes, including elements from internet and popular culture as well as socio-cultural narratives. However, gender-sensitive and neutral topic choices enhance engagement, self-efficacy, contribution, code quality and creative output, while also contributing to reduce stereotypical beliefs about programming, particularly among girls. In line with the findings for the course topic, group constellations also influence programming experiences. In particular, introducing pair programming in courses shows a promising approach for young learners, but attention must be paid to mitigate socially learned gender-stereotypical behaviours. Another finding indicates that, unlike professional software teams, mixed-diverse student teams often encounter substantial challenges, thus benefit from clear communication guidelines and supportive environments to promote better collaboration. This doctoral thesis concludes with guidelines for designing more effective and inclusive introductory programming courses. These recommendations include using gender-sensitive course materials, allowing for creative freedom through topic choices while encouraging the use of advanced programming concepts, promoting collaboration through pair programming while fostering enhanced communication, boosting self-efficacy with quick positive feedback for girls in particular, and providing emotional support for underrepresented groups. By following these guidelines, educators can create more engaging, inclusive, and effective programming courses. This may ultimately promote a more equitable and diverse future generation of professional software developers while also fostering computational thinking, encouraging a broader interest in programming among all young learners. KW - Softwareentwicklung KW - Lernsituation Y1 - 2024 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-15049 ER - TY - THES A1 - Berger, Christian T1 - Towards Fast and Adaptive Byzantine State Machine Replication for Planetary-Scale Systems N2 - State machine replication (SMR) is a classical approach for building resilient distributed systems. In Byzantine fault-tolerant (BFT) systems, no concrete assumptions are made about the behavior of faulty replicas. With the advancement of distributed ledger technologies (DLT), planetary-scale BFT SMR ist becoming practical and necessary as it can serve as a consensus primitive to keep the ledger consistent. In our view, the alignment of BFT SMR to DLT brings new challenges, for instance the scalability aspect, where recent research works less frequently address latency improvements than throughput improvements. Further challenges include the geographic dispersion of replicas within a planetary-scale system and the need of a BFT SMR protocol to react to environmental changes during runtime. This thesis has the objective to improve BFT SMR for planetary-scale systems by lowering the protocol latency observed by clients and by making the BFT SMR system adaptive, i.e., enabling replicas to react to perceived changes such as changing network characteristics or faulty replicas. As a first contribution of this thesis, we discover that fast, consensus-free (read-only) operations is a flawed optimization in seminal BFT SMR frameworks, such as PBFT and BFT-SMaRt. We explain how the read-only optimization can violate the protocol's liveness by showing an attack and then present a solution that makes the overall, optimized protocol both live and linearizable. The second contribution is Adaptive Wide-Area Replication (AWARE), which enables a geo-replicated system to adapt to its environment, thus improving the geographical scalability of consensus if replicas are dispersed across the world. Essentially, AWARE is an automated and dynamic voting-weight tuning and leader positioning scheme, which supports the emergence of fast consensus quorums in the system and builds upon previous work, the WHEAT protocol. AWARE combines reliable self-monitoring with a consensus latency prediction model, thus striving to minimize the system’s consensus latency at runtime, which subsequently results in latency improvements observed by clients scattered across the globe, which we validate through experiments. The third contribution presents FlashConsensus, a protocol derived from AWARE, that also adjusts the resilience threshold. The core idea is the tentative use of a lower resilience threshold which leads to smaller consensus quorums and thus consensus acceleration in common-case scenarios where we expect only few faulty replicas. FlashConsensus achieves threat-level awareness through the incorporation of two modes of operation and BFT forensic support and guarantees liveness and linearizability under optimal resilience. Moreover, FlashConsensus allows for client-side speculation by using incremental consistency guarantees to further lower request latency. Additionally, we investigate on the question whether we can reason about the performance of large-scale systems utilizing simulations. We discover, that we can faithfully forecast the performance of BFT protocols by plugging real protocol implementations into a high-performance network simulator. For instance, simulation results reveal that, using 51 replicas scattered across the planet, FlashConsensus can finalize operations in less than 0.4 s, which is half of the time required for a PBFT-like protocol in the same network, and matching the latency of this protocol running on the best possible internet links (transmitting at 67% of the speed of light). N2 - Die Zustandsmaschinenreplikation (ZMR) ist ein klassischer Ansatz für zuverlässige verteilte Systeme. In Byzantinischen fehlertoleranten (BFT) Systemen werden keine konkreten Annahmen über das Verhalten von fehlerhaften Replikaten gemacht. Mit dem Voranschreiten der sog. „Distributed Ledger Technologien“ (DLT) wird planetare BFT ZMR praktikabel und notwendig, da sie als Konsensusprimitiv dienen kann, um die Konsistenz einer Blockchain aufrechtzuerhalten. Unserer Ansicht nach bringt die Ausrichtung von BFT ZMR an DLT neue Herausforderungen mit sich, z.B. den Skalierbarkeitsaspekt, bei dem jüngste Forschungsarbeiten seltener Latenzverbesserungen als Durchsatzverbesserungen untersuchen. Weitere Herausforderungen umfassen die geografische Verteilung von Replikaten innerhalb eines planetaren Systems sowie die Notwendigkeit eines BFT ZMR Protokolls während der Laufzeit auf Veränderungen zu reagieren. Diese Dissertation verfolgt das Ziel, die BFT ZMR für planetare Systeme durch Senkung der von Clients beobachteten Protokolllatenz zu verbessern und die BFT ZMR-Systeme anpassungsfähig zu machen, indem Replikate auf wahrgenommene Änderungen wie sich ändernde Netzwerkcharakteristiken oder fehlerhafte Replikate reagieren können. Als erster Beitrag dieser Dissertation zeigen wir, dass schnelle, konsensusfreie (nur-lesende) Operationen in grundlegenden BFT ZMR Protokollen, wie PBFT und BFT-SMaRt, eine fehlerhafte Optimierung sind. Wir erklären, wie die nur-lesende Optimierung die Liveness (Verfügbarkeit von Operationen) des Protokolls verletzen kann, indem wir einen Angriff präsentieren und dann eine Lösung vorschlagen, die das insgesamt optimierte Protokoll sowohl live (verfügbar) als auch linearisierbar („linearizable“ -- streng konsistent) macht. Der zweite Beitrag ist Adaptive Wide-Area Replication (AWARE), mit der ein geo-repliziertes System sich an seine Umgebung anpassen kann und damit die geografische Skalierbarkeit des Konsensus verbessert, wenn Replikate auf der ganzen Welt verteilt sind. Im Wesentlichen ist AWARE ein automatisches und dynamisches Stimmgewichtseinstellungs- und Anführerpositionierungs-schema, das die Entstehung schneller Konsensusquoren im System unterstützt und auf früheren Arbeiten, wie dem WHEAT-Protokoll, aufbaut. AWARE kombiniert zuverlässige Selbstüberwachung mit einem Konsensuslatenzvorhersagemodell und strebt danach, die Konsensuslatenz des Systems zur Laufzeit zu minimieren, was letztendlich zu Latenzgewinnen führt, die von Clients auf der ganzen Welt beobachtbar sind, was wir durch Experimente bestätigen. Der dritte Beitrag stellt FlashConsensus vor, ein Protokoll, das aus AWARE abgeleitet ist und auch die Resilienzschwelle anpasst. Die Kernidee ist die vorübergehende Verwendung einer niedrigeren Resilienzschwelle, die zu kleineren Konsensusquoren und damit zu einer Beschleunigung des Konsensus in häufigen Fällen führt, in denen nur wenige fehlerhafte Replikate erwartet werden. FlashConsensus erkennt und reagiert auf Bedrohungen durch die Einbeziehung von zwei Betriebsarten und BFT-Forensikunterstützung und garantiert Liveness sowie Konsistenz unter optimaler Resilienz. Darüber hinaus ermöglicht es die Spekulation auf Clientseite durch die Verwendung inkrementeller Konsistenzgarantien, um die Clientlatenz weiter zu senken. Zusätzlich werden wir uns mit der Frage beschäftigen, ob wir die Performanz von groß-skalierten Systemen mittels Simulationen beurteilen können. Wir stellen fest, dass wir die Performanz von BFT Protokollen durch das Einstöpseln von echten Protokollimplementierungen in einen hochleistungsfähigen Netzwerksimulator zuverlässig vorhersagen können. Beispielsweise zeigen Simulationen, dass FlashConsensus mit 51 Replikaten, die auf der ganzen Welt verteilt sind, Operationen in weniger als 0,4 s linearisierbar verarbeiten kann, was die Hälfte der Zeit ist, die für ein PBFT-ähnliches Protokoll im gleichen Netzwerk erforderlich ist und ähnlich schnell ist wie dieses Protokoll mit bestmöglichen Internetverbindungen (Übertragung mit 67% der Lichtgeschwindigkeit). KW - Byzantine fault tolerance KW - state machine replication KW - consensus KW - adaptiveness KW - planetary-scale Y1 - 2024 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-15059 ER - TY - THES A1 - Sentanoe, Stewart T1 - VMIaaS: Virtual Machine Introspection as a Service N2 - In this digital era, communication in the digital world is becoming part of our daily lives. One key technology that becomes part of this digital transformation is cloud computing. It allows users to have a running system as a virtual machine (VM) on the cloud without owning a physical server. Unfortunately, adversaries can also use those systems to conduct criminal activities. Therefore, developing a method to extract evidence from those systems is also necessary. One way is through digital forensics, and one method to do digital forensics of a VM is using virtual machine introspection (VMI). However, VMI has yet to be made available by any public cloud provider. This thesis addresses this issue by introducing methods for deploying VMI on public cloud providers. Four main challenges have to be solved. Firstly, VMI requires access to the hypervisor, which practically can access all VMs running on the same server. This leads to security and privacy issues where customers can introspect each other VMs. To solve this problem, this thesis introduces KVMIveggur, a versatile access control of VMI. It comes with different options that every customer can choose from based on their needs. Secondly, VMI introduces overhead to the running VM. This is because most of the introspection mechanisms perform data access to the monitored VM. Performing data access on a running VM can cause data inconsistency. Hence, pausing the VM before executing the data access is better. However, when the VM pausing frequency is high, it will affect the performance of the monitored VM. The current state-of-the-art techniques use caching to reduce the VM pausing frequency. However, it faces a problem: the cached data may be outdated compared to the actual data. Therefore, this thesis introduces VMIFresh, a better caching mechanism. We leverage both active and passive tracing mechanisms to ensure high performance and consistency of the data (freshness). Thirdly, many state-of-the-art VMI libraries and applications run perfectly only on Intel processors because Intel CPUs provide the best hardware support for VMI. However, AMD and ARM processors are getting more popular in cloud computing. Thus, it is necessary to retrofit VMI capabilities to support AMD and ARM processors. This thesis describes the requirements to employ VMI on AMD and ARM processors. We also provide the implementation of those requirements. Finally, to do introspection using VMI, it is crucial to have proper symbol information (layout and location of data structures) of the introspected operating system (OS) and user applications. While many existing VMI approaches concentrate primarily on analyzing OS data structures, analyzing user application data often receives no attention. In our approach, we address this gap by focussing on application-level introspection. We have identified several use cases that require this kind of introspection. We focus on cryptographic key extraction for two specific instances: secure shell (SSH) and transport layer security (TLS) by leveraging the power of machine learning techniques to locate those keys in the main memory effectively and efficiently. After we had solved those challenges, we combined a couple of our approaches and introduced two VMI applications: Sarracenia and VMIGuard. Sarracenia is a deception technology that tracks activities done on an SSH session. The main goal of Sarracenia is to attract adversaries away from the production system and learn about their behavior. On the other hand, VMIGuard also monitors the SSH traffic. But, it specifically monitors the activity of any git-related activities. The main goal of VMIGuard is to ensure the integrity of the hosted data from any internal malicious actor. KW - Cloud Computing KW - Computersicherheit Y1 - 2024 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:739-opus4-15027 ER -