Augmented reality applications are bitrate intensive, delay-sensitive, and computationally demanding. To support them, mobile edge computing systems need to carefully manage both their networking and computing resources. To this end, we present a proof of concept resource management scheme that adapts the bandwidth at the base station and the GPU frequency at the edge to efficiently fulfill roundtrip delay constrains. Resource adaptation is performed using a Multi-Armed Bandit algorithm that accounts for the monotonic relationship between allocated resources and performance. We evaluate our scheme by experimentation on an OpenAirInterface 5G testbed where the considered application is OpenRTiST. The results indicate that our resource management scheme can substantially reduce both bandwidth usage and power consumption while delivering high quality of service. Overall, this work demonstrates that intelligent resource control can potentially establish systems that are not only more efficient but also more sustainable.

With the emergence of time-critical applications in modern communication networks such as vehicle-to-everything (V2X) systems, there is a growing demand for proactive network adaptation and quality of service (QoS) prediction. However, a fundamental question remains largely unexplored: How can we quantify and achieve more predictable communication systems in terms of performance? To address this gap, this paper introduces a theoretical framework for defining and analyzing predictability in communication systems, with a focus on the impact of observations for performance forecasting. We establish a mathematical definition of predictability based on the total variation distance between the forecast and marginal performance distributions. A system is deemed unpredictable when the forecast distribution, providing the most comprehensive characterization of future states using all accessible information, is indistinguishable from the marginal distribution, which depicts the system's behavior without any observational input. This framework is applied to multi-hop systems under Markovian conditions, with a detailed analysis of Geo/Geo/1/K queuing models in both single-hop and multi-hop scenarios. Additionally, we apply the framework to a random-walk-based model of QoS for connected vehicles experiencing changing channel conditions. We derive exact and approximate expressions for predictability in these systems, as well as upper bounds based on spectral analysis of the underlying Markov chains. Our results have implications for the design of efficient monitoring and prediction mechanisms in future communication networks aiming to provide dependable services.

The explosive growth of mobile communication and the proliferation of real-time applications, such as industrial automation and extended reality (XR), have created unprecedented demands for ultra-reliable low-latency communication (URLLC) in wireless networks. For example, in industrial closed-loop control systems, data must be transmitted within a target delay of atmost a few milliseconds; violations can lead to costly failures and, there-fore, must occur with probabilities below 0.0001 (or, reliability above 0.9999).This dissertation addresses the critical challenge of end-to-end latency pre-diction and control in these dynamic and stochastic environments, bridging the gap between the inherent randomness of wireless communication and the deterministic performance guarantees required by time-sensitive applications.

In this thesis, we adopt a twofold approach, combining rigorous theoretical analysis with practical, data-driven methodologies. First, we introduce a framework for analyzing predictability that quantifies the inherent limits of latency forecasting in communication networks. Through analysis of Marko-vian systems, including single-hop and multi-hop queues, exact expressions and spectral-based upper bounds for predictability are derived, revealing the crucial influence of network topology, state transitions, and observation defects. Building on this foundation, we developed and implemented data-driventechniques for probabilistic delay prediction. A key contribution is a tail-optimized prediction method that integrates Extreme Value Theory (EVT) within a mixture density network framework, significantly enhancing the accuracy of predicting rare, high-latency events critical for URLLC. To demonstrate the practical utility of these predictions, ”Delta,” a novel active queue management scheme, is introduced. Delta integrates real-time delay violation probability predictions into packet-dropping decisions, dynamically adapting to delay variations and significantly reducing delay violations. To validate these approaches, the ExPECA testbed and EDAF framework were developed, enabling fine-grained delay measurement and decomposition in real 5G systems. Extensive experiments on both commercial off-the-shelf5G and software-defined radio-based Open Air Interface platforms confirmedthe superior accuracy and efficiency of the proposed EVT-enhanced models.

Furthermore, temporal prediction models, leveraging LSTM and Transformer architectures, were developed and shown to achieve higher accuracy comparedto the baseline approaches in real 5G network experiments, capturing the time-varying dynamics of wireless networks and providing accurate multi-step forecasts. This dissertation advances latency prediction and control for wireless networks, offering both theoretical foundations and practical solutions for time-sensitive applications. These findings have significant implications for designing and operating next-generation wireless networks, paving the way for more dependable communication. Future work should focus on integrating these prediction models to optimize the network and extending the framework to encompass broader quality of service metrics and emerging wireless technologies.  

Den explosionsartade tillväxten av mobil kommunikation och spridningen av realtidsapplikationer, såsom industriell automation och utökad verklighet (XR), har skapat enastående krav på ultratillförlitlig kommunikation med låg fördröjning (URLLC) i trådlösa nätverk. Till exempel måste data i industriella slutna styrsystem ¨överföras inom en deadline på högst några millisekunder; ¨överträdelser kan leda till kostsamma fel och måste därför inträffa med sannolikheter under 0,0001 (eller, en tillförlitlighet över 0,9999). Denna avhandling behandlar den kritiska utmaningen att prediktera och kontrollera fördröjningen mellan sändare till mottagare i dessa dynamiska och stokastiska miljöer, och minskar skillnaden mellan den inneboende slumpmässigheten i trådlös kommunikation och de deterministiska prestandagarantier som krävs av tidskänsliga applikationer. I denna avhandling antas en tvådelad metod som kombinerar noggrann teoretisk analys med praktiska, datadrivna metoder. Först introduceras ett ramverk för att analysera förutsägbarhet som kvantifierar de inneboende gränserna för fördröjningsprognoser i kommunikationsnätverk. Genom att studera Markovsystem, däribland enkel- och multihoppköer, härleds exakta uttryck och spektrumbaserade övre gränser för förutsägbarhet, vilket belyser hur nätverkstopologi, tillståndsövergångar och observationsdefekter påverkar resultaten.

Utifrån denna grund utvecklades och implementerades datadrivna tekniker för probabilistisk fördröjningsprediktion. Ett viktigt bidrag är en metod för prediktion som integrerar extremvärdesteori (EVT) i ett ramverk för blandningstäthetsnätverk, vilket avsevärt förbättrar förmågan att prediktera sällsynta, höga fördröjningar som är avgörande för URLLC. För att demonstrera den praktiska nyttan av dessa prediktioner presenteras ”Delta,”ett nytt aktivt köhanteringssystem. Delta integrerar, i realtid, prediktioner av sannolikheten för fördröjningsöverträdelser i beslutsprocessen för paketborttagning, vilket minskar fördröjningsöverträdelser avsevärt.

För att validera dessa metoder utvecklades testbädden ExPECA och ramverket EDAF, som möjliggör högupplösta mätningar och uppdelning av fördröjningens komponenter i verkliga 5G-system. Omfattande experiment på både kommersiell 5G-utrustning och mjukvarudefinierade radioplattformar baserade på Open Air Interface bekräftade den förbättrade noggrannheten och effektiviteten hos de föreslagna EVT-förbättrade modellerna. Vidare utvecklades temporala prediktionsmodeller som använder LSTM- och Transformer-arkitekturer som visade högre träffsäkerhet än referensmetoder i verkliga 5G-nätverksexperiment, då de fångar de tidsvarierande dynamikerna i trådlösa nätverk och möjliggör exakta flerstegsprognoser.

Denna avhandling driver framåt forskningen om fördröjningsprediktion och -kontroll i trådlösa nätverk och erbjuder både teoretiska grunder och praktiska lösningar för tidskänsliga applikationer. Resultaten har stor betydelse för utformningen och driften av nästa generations trådlösa nätverk och banar väg för mer pålitlig kommunikation. Framtida arbete ska/borde/kan (will/should/can) fokusera på att integrera dessa prediktionsmodeller för att optimera nätverket, och utvidga ramverket till att omfatta bredare kvalitetsmätningar och nya trådlösa teknologier.

Supporting applications in emerging domains like cyber-physical systems and human-in-the-loop scenarios typically requires adherence to strict end-to-end delay guarantees. Contributions of many tandem processes unfolding layer by layer within the wireless network result in violations of delay constraints, thereby severely degrading application performance. Meeting the application's stringent requirements necessitates coordinated optimization of the end-to-end delay by fine-tuning all contributing processes. To achieve this task, we designed and implemented EDAF, a framework to decompose packets' end-to-end delays and determine each component's significance for 5G network. We showcase EDAF on OpenAirInterface 5G uplink, modified to report timestamps across the data plane. By applying the obtained insights, we optimized end-to-end uplink delay by eliminating segmentation and frame-alignment delays, decreasing average delay from 12ms to 4ms.

The increasing demand for latency-sensitive applications has necessitated the development of sophisticated algorithms that efficiently manage packets with end-to-end delay targets traversing the networked infrastructure. Network components must consider minimizing the packets' end-to-end delay violation probabilities (DVP) as a guiding principle throughout the transmission path to ensure timely deliveries. Active queue management (AQM) schemes are commonly used to mitigate congestion by dropping packets and controlling queuing delay. Today's established AQM schemes are threshold-driven, identifying congestion and trigger packet dropping using a predefined criteria which is unaware of packets' DVPs. In this work, we propose a novel framework, Delta, that combines end-to-end delay characterization with AQM for minimizing DVP. In a queuing theoretic environment, we show that such a policy is feasible by utilizing a data-driven approach to predict the queued packets' DVPs. That enables Delta AQM to effectively handle links with arbitrary stationary service time processes. The implementation is described in detail, and its performance is evaluated and compared with state of the art AQM algorithms. Our results show the Delta outperforms current AQM schemes substantially, in particular in scenarios where high reliability, i.e. high quantiles of the tail latency distribution, are of interest.

With the emergence of new application areas, such as cyber-physical systems and human-in-the-loop applications, there is a need to guarantee a certain level of end-to-end network latency with extremely high reliability, e.g., 99.999%. While mechanisms specified under IEEE 802.1AS time-sensitive networking (TSN) can be used to achieve these requirements for switched Ethernet networks, implementing TSN mechanisms in wireless networks is challenging due to their stochastic nature. To conform the wireless link to a reliability level of 99.999%, the behavior of extremely rare outliers in the latency probability distribution, or the tail of the distribution, must be analyzed and controlled. This work proposes predicting the tail of the latency distribution using state-of-the-art data-driven approaches, such as mixture density networks (MDN) and extreme value mixture models, to estimate the likelihood of rare latencies conditioned on the network parameters, which can be used to make more informed decisions in wireless transmission. Actual latency measurements of a commercial private and a software-defined 5G network are used to benchmark the proposed approaches and evaluate their sensitivities concerning the tail probabilities. Our benchmarks highlight how the proposed methods, aided by noise regularization, achieve an acceptable accuracy in the extreme 99.9999% latency probabilities.

This paper presents ExPECA, an edge computing and wireless communication research testbed designed to tackle two pressing challenges: comprehensive end-to-end experimentation and high levels of experimental reproducibility. Leveraging OpenStack-based Chameleon Infrastructure (CHI) framework for its proven flexibility and ease of operation, ExPECA is located in a unique, isolated underground facility, providing a highly controlled setting for wireless experiments. The testbed is engineered to facilitate integrated studies of both communication and computation, offering a diverse array of Software-Defined Radios (SDR) and Commercial Off-The-Shelf (COTS) wireless and wired links, as well as containerized computational environments. We exemplify the experimental possibilities of the testbed using OpenRTiST, a latency-sensitive, bandwidthintensive application, and analyze its performance. Lastly, we highlight an array of research domains and experimental setups that stand to gain from ExPECA's features, including closed-loop applications and time-sensitive networking.

End-to-end learning for wireless communications has recently attracted much interest in the community, owing to the emergence of deep learning-based architectures for the physical layer. Neural network-based autoencoders have been proposed as potential replacements of traditional model-based transmitter and receiver structures. Such a replacement primarily provides an unprecedented level of flexibility, allowing to tune such emerging physical layer network stacks in many different directions. The semantic relevance of the transmitted messages is one of those directions. In this paper, we leverage a specific semantic relationship between the occurrence of a message (the source), and the channel statistics. Such a scenario could be illustrated for instance, in vehicular communications where the distance is to be conveyed between a leader and a follower. We study two autoencoder approaches where these special circumstances are exploited. We then evaluate our autoencoders, showing through the simulations that the semantic optimization can achieve significant improvements in the BLERs (up till 93.6%) and RMSEs (up till 87.3%) for vehicular communications leading to considerably reduced risks and needs for message retransmissions.

Experimental research on wireless networking in combination with edge and cloud computing has been the subject of explosive interest in the last decade. This development has been driven by the increasing complexity of modern wireless technologies and the extensive softwarization of these through projects such as a Open Radio Access Network (O-RAN). In this context, a number of small- to mid-scale testbeds have emerged, employing a variety of technologies to target a wide array of use-cases and scenarios in the context of novel mobile communication technologies such as 5G and beyond-5G. Little work, however, has yet been devoted to developing a standard framework for wireless testbed automation which is hardwareagnostic and compatible with edge- and cloud-native technologies. Such a solution would simplify the development of new testbeds by completely or partially removing the requirement for custom management and orchestration software. In this paper, we present the first such mostly hardwareagnostic wireless testbed automation framework, Ainur. It is designed to configure, manage, orchestrate, and deploy workloads from an end-to-end perspective. Ainur is built on top of cloudnative technologies such as Docker, and is provided as FOSS to the community through the KTH-EXPECA/Ainur repository on GitHub. We demonstrate the utility of the platform with a series of scenarios, showcasing in particular its flexibility with respect to physical link definition, computation placement, and automation of arbitrarily complex experimental scenarios.

With the advent of edge computing, there is increasing interest in wireless latency-critical services. Such applications require the end-to-end delay of the network infrastructure (communication and computation) to be less than a target delay with a certain probability, e.g., 10(-2)-10(-5). To deal with this guarantee level, the first step is to predict the transient delay violation probability (DVP) of the packets traversing the network. The guarantee level puts a threshold on the tail of the end-to-end delay distribution; thus, it makes data-driven DVP prediction a challenging task. We propose to use the extreme value mixture model in the mixture density network (MDN) method for this task. We implemented it in a multi-hop queuing-theoretic system to predict the DVP of each packet from the network state variables. This work is a first step toward utilizing the DVP predictions, possibly in the resource allocation scheme or queuing discipline. Numerically, we show that our proposed approach outperforms state-of-the-art Gaussian mixture model-based predictors by orders of magnitude, in particular for scenarios with guarantee levels above 10(-2).