Our world is becoming more and more connected. The Internet of things (IoT) is one of the emerging technologies connecting an increasing number of devices to the Internet and offering value-added services. The desire for more and more connectivity and smart control, the low technical effort, and the decreased cost of Internet-enabled devices favor this trend. IoT enriches a variety of application domains, including home automation, smart grid, Industrial Internet of Things (IIoT), vehicle-to-everything (V2X) communication, smart city, health care, disaster forecasting, and the observation of animals and nature. A high diversity of IoT devices ranges from tiny battery-powered sensors to artificial intelligence (AI) applications running on powerful hardware. The types of data transmission are also heterogeneous, such as rarely transmitted sensor data of just a few bytes, high-frequency messaging data, or large, continuous data streams. The dynamics of IoT systems, the amount of IoT devices, and the high diversity introduce challenges to applications and infrastructure. The infrastructure components include computing nodes, backend software systems, databases, storage systems, as well as access and data center networks. The systems must consider performance aspects in order to maintain the desired quality of experience (QoE) for users. Due to varying requirements, the systems need to consider performance aspects both during design as well as during runtime by supporting adaptation. This dissertation deals with design optimizations and runtime adaptations of networks from the perspective of application developers and data center operators through measurement, modeling, and prediction of network quality with a focus on IoT. This work covers the three research areas of: (i) quality prediction in mobile networks, (ii) performance evaluation and testing of communication protocols, and (iii) adaptation of data center networks. The first research area addresses quality prediction in mobile networks. Applications must react to changing network conditions, such as varying coverage or even dead spots, which are intrinsic to mobile networks. Such adjustments include, for example, the adjustment of a video bitrate or the deferral of low-priority data transfers. A long-term network quality prediction is beneficial for such communication adaptation mechanisms. Furthermore, applications could inform users in advance about areas with limited network coverage, and route planning software can include the prediction. However, existing approaches focus primarily on short-term predictions and require current connectivity parameters. We identified the need for prediction approaches that provide long-term predictions with a horizon of several minutes or even hours. The second research area deals with the performance evaluation and testing of communication protocols. Various communication protocols have emerged over time, differing in their suitability for specific scenarios. IoT system designers and application developers should be aware of the characteristics of the communication protocols and consider them when selecting a suitable protocol. Existing work on performance evaluation of communication protocols focus on scalability and efficiency in stable networks. This work addresses the shortcomings of such tools by supporting the performance evaluation of communication protocols under constrained and varying network quality conditions. The third research area concerns the adaptation of data center networks. The varying number of devices and load also affects the data center networks. Service providers that run their applications in the cloud define the requirements on infrastructure performance through service-level agreements (SLAs) with the data center operators. Data center operators must continuously monitor their networks for SLA-compliance and respond to performance bottlenecks. Existing network adaptation approaches focus on specific adaptation aspects and technologies such as rerouting or software-defined networking (SDN) operations. This work identifies the need for an approach supporting different, technology-agnostic network adaptation strategies. Such an approach should suggest adaptations to the data center operators or apply them directly in order to cope with the dynamics of communication. The adaptation approach should validate the adaptation options automatically before suggesting them. Motivated by the identified shortcomings and needs, we summarize the contributions of this dissertation as follows. Technique for continuously measuring cellular network quality: We design and implement a technique for collecting measurements in cellular networks with the following key characteristics. First, the measurement technique supports continuous and automatic measurements, enabling car drivers or cyclists to collect measurements without any user interaction. Second, the technique integrates a fast bitrate measurement algorithm with low data volume consumption, making it well-suited for measurements within short time intervals and data collection by volunteers. Third, we implemented the technique on the Android operating system, eliminating the need for specialized measurement equipment, which facilitates capturing measurements by a crowd of users using commodity smartphones. Collecting, processing, and publishing a cellular network measurement dataset: Using our developed technique for quality measurements in cellular networks, we collected 326,157 measurements and published them as open data. We captured the measurements primarily in Germany within a small region. Repeated use of the same routes allows investigation of network quality changes over time and under different conditions. This dissertation analyzes the dataset and provides algorithms for detecting implausible and abnormal values. The cleaned dataset is particularly suitable for further processing in research, for example, as input for machine learning algorithms. Prediction models for connection quality in cellular networks: We developed an approach for predicting the download bitrate in cellular networks along a route. The approach uses a machine learning model trained on the collected dataset. The model requires passive measurements along the route for a prediction. These measurements do not need to send or receive any packets and can, therefore, be collected by a crowd of users through an application in the background. By eliminating the need for actual connection parameters, our approach allows prediction well in advance and what-if analysis for long-term decisions. The prediction enables applications to react in advance or inform the user about the expected network quality and coverage. Benchmarking framework for the performance analysis of publish/subscribe protocols: We developed ComBench, a benchmarking framework for publish/subscribe protocols, focusing on networks with limited and varying quality. ComBench emulates clients that send and receive messages according to a workload definition. These clients can switch between different publish/subscribe protocols. During the benchmark run, ComBench can apply constraints like bitrate limitations or packet losses on their network connection. ComBench provides built-in collection and reporting of performance metrics. IoT system designers can use ComBench to test IoT scenarios matching real-world conditions in wireless networks. Network emulator for testing the communication behavior of applications: Our developed IoT network emulator can emulate networks with constrained and varying network conditions. Software developers can easily integrate it into software testing pipelines through its simple instantiation via the command line and its provided interfaces. These pipelines can then automatically test the desired behavior of an application under different network conditions for each new release. Online adaptation framework for data center networks: We designed an online adaptation framework for data center networks that utilizes a network model to simulate the workload for detecting bottlenecks and SLA violations. A control loop applies adaptation strategies on the model, simulates the suggested configuration changes, and iteratively executes additional adaptation strategies as needed. The validation through simulation and automated optimization ensures that the framework only suggests SLA compliant and cost-optimal configurations. Our developed approaches are valuable contributions to academic research as well as practical applications. To the best of our knowledge, our work is the first to provide a long-term prediction approach for mobile network quality without extensive prior bitrate measurements. Evaluation of the approach shows that it can estimate connection quality trends well. Regarding the performance evaluation of application layer protocols, our approaches differ from existing tools through their multi-protocol-support and capabilities for emulating network limitations and quality variations. A case study illustrates the different application domains of our performance evaluation approaches. The main characteristics of our approach for the online adaptation of data center networks are technology independence and validation of the suggested configuration changes. The evaluation shows that the adaptation algorithm scales well even for more extensive networks and is Pareto-optimal with respect to multiple cost dimensions.

The steadily increasing usage of smart meters generates a valuable amount of high-resolution data about the individual energy consumption and production of local energy systems. Private households install more and more photovoltaic systems, battery storage and big consumers like heat pumps. Thus, our vision is to augment these collected smart meter time series of a complete system (e.g., a city, town or complex institutions like airports) with simulatively added previously named components. We, therefore, propose a novel digital twin of such an energy system based solely on a complete set of smart meter data including additional building data. Based on the additional geospatial data, the twin is intended to represent the addition of the abovementioned components as realistically as possible. Outputs of the twin can be used as a decision support for either system operators where to strengthen the system or for individual households where and how to install photovoltaic systems and batteries. Meanwhile, the first local energy system operators had such smart meter data of almost all residential consumers for several years. We acquire those of an exemplary operator and discuss a case study presenting some features of our digital twin and highlighting the value of the combination of smart meter and geospatial data.

Scalability is often mentioned in literature, but a stringent definition is missing. In particular, there is no general scalability assessment which clearly indicates whether a system scales or not or whether a system scales better than another. The key contribution of this article is the definition of a scalability index (SI) which quantifies if a system scales in comparison to another system, a hypothetical system, e.g., linear system, or the theoretically optimal system. The suggested SI generalizes different metrics from literature, which are specialized cases of our SI. The primary target of our scalability framework is, however, benchmarking of two systems, which does not require any reference system. The SI is demonstrated and evaluated for different use cases, that are (1) the performance of an IoT load balancer depending on the system load, (2) the availability of a communication system depending on the size and structure of the network, (3) scalability comparison of different location selection mechanisms in fog computing with respect to delays and energy consumption; (4) comparison of time-sensitive networking (TSN) mechanisms in terms of efficiency and utilization. Finally, we discuss how to use and how not to use the SI and give recommendations and guidelines in practice. To the best of our knowledge, this is the first work which provides a general SI for the comparison and benchmarking of systems, which is the primary target of our scalability analysis.

Background: Due to the importance of radiologic examinations, such as X-rays or computed tomography scans, for many clinical diagnoses, the optimal use of the radiology department is 1 of the primary goals of many hospitals. Objective: This study aims to calculate the key metrics of this use by creating a radiology data warehouse solution, where data from radiology information systems (RISs) can be imported and then queried using a query language as well as a graphical user interface (GUI). Methods: Using a simple configuration file, the developed system allowed for the processing of radiology data exported from any kind of RIS into a Microsoft Excel, comma-separated value (CSV), or JavaScript Object Notation (JSON) file. These data were then imported into a clinical data warehouse. Additional values based on the radiology data were calculated during this import process by implementing 1 of several provided interfaces. Afterward, the query language and GUI of the data warehouse were used to configure and calculate reports on these data. For the most common types of requested reports, a web interface was created to view their numbers as graphics. Results: The tool was successfully tested with the data of 4 different German hospitals from 2018 to 2021, with a total of 1,436,111 examinations. The user feedback was good, since all their queries could be answered if the available data were sufficient. The initial processing of the radiology data for using them with the clinical data warehouse took (depending on the amount of data provided by each hospital) between 7 minutes and 1 hour 11 minutes. Calculating 3 reports of different complexities on the data of each hospital was possible in 1-3 seconds for reports with up to 200 individual calculations and in up to 1.5 minutes for reports with up to 8200 individual calculations. Conclusions: A system was developed with the main advantage of being generic concerning the export of different RISs as well as concerning the configuration of queries for various reports. The queries could be configured easily using the GUI of the data warehouse, and their results could be exported into the standard formats Excel and CSV for further processing.

Cooperative, connected and automated mobility (CCAM) systems depend on a reliable communication to provide their service and more crucially to ensure the safety of users. One way to ensure the reliability of a data transmission is to use multiple transmission technologies in combination with redundant flows. In this paper, we describe a system requiring multipath communication in the context of CCAM. To this end, we introduce a data plane-based scheduler that uses replication and integration modules to provide redundant and transparent multipath communication. We provide an analytical model for the full replication module of the system and give an overview of how and where the data-plane scheduler components can be realized.