We optimize hierarchies of Time-to-Live (TTL) caches under network delays. A TTL cache assigns individual eviction timers to cached objects that are usually refreshed upon a hit where upon a miss the object requires a random time to be fetched from a parent cache. Due to their object decoupling property, TTL caches are of particular interest since the optimization of a per-object utility enables service differentiation. However, state-of-the-art exact TTL cache utility-based optimization does not extend beyond single TTL caches, especially under network delays. In this paper, we leverage the object decoupling effect to formulate the nonlinear utility maximization problem for TTL cache hierarchies in terms of the exact object hit probability under random network delays. We iteratively solve the utility maximization problem to find the optimal per-object TTLs. In addition, we propose a variant TTL policy, which we denote as exTTL to counteract the effect on the optimal utility of the storage mismatch between the actual realization of a TTL cache and its ideal infinite storage assumption. Further, we show that the exact model suffers from tractability issues for large hierarchies and propose a machine learning approach to estimate the optimal TTL values for large systems. Finally, we provide numerical and data center trace-based evaluations for both methods, showing the significant offloading improvement due to TTL optimization considering the network delays.

Performance modeling is crucial for managing wireless communication systems, offering insights into complex networks and enabling optimization. In recent years, Machine Learning algorithms have become indispensable for modeling both static and vehicular networks. Among these, Graph Neural Networks have gained prominence due to their flexibility and ability to capture intricate network structures. While Graph Neural Networks have been successfully applied to predict Key Performance Indicators such as delay in multi-hop networks, their effectiveness in Airborne Networks remains underexplored. In this study, we investigate the applicability of Graph Neural Network algorithms, originally developed for static networks, in modeling the performance of Airborne Networks. Our findings reveal that performance of Graph Neural Network in Airborne Networks falls short of their effectiveness in static environments. To address this, we introduce adaptations through enhanced input features, improving their generalization to high network velocity levels and increasing overall performance.

The integration of unmanned aerial vehicles (UAVs) into cellular networks presents significant mobility management challenges, primarily due to frequent handovers caused by probabilistic line-of-sight conditions with multiple ground base stations (BSs). To tackle these challenges, reinforcement learning (RL)-based methods, particularly deep Q-networks (DQN), have been employed to optimize handover decisions dynamically. However, a major drawback of these learning-based approaches is their black box nature, which limits interpretability in the decision making process. This paper introduces an explainable AI (XAI) framework that incorporates Shapley Additive Explanations (SHAP) to provide deeper insights into how various state parameters influence handover decisions in a DQN-based mobility management system. By quantifying the impact of key features such as reference signal received power (RSRP), reference signal received quality (RSRQ), buffer status, and UAV position, our approach enhances the interpretability and reliability of RL-based handover solutions. To validate and compare our framework, we utilize real-world network performance data collected from UAV flight trials. Simulation results show that our method provides intuitive explanations for policy decisions, effectively bridging the gap between AI-driven models and human decision-makers.

The derivation of upper bounds on data flows’ worst-case traversal times is an important task in many application areas. For accurate bounds, model simplifications should be avoided even in large networks. Network Calculus (NC) provides a modeling framework and different analyses for delay bounding. We investigate the analysis of feedforward networks where all queues implement First-In First-Out (FIFO) service. Correctly considering the effect of data flows onto each other under FIFO is already a challenging task. Yet, the fastest available NC FIFO analysis (called LUDB) suffers from limitations resulting in unnecessarily loose bounds. A feature called Flow Prolongation (FP) has been shown to improve delay bound accuracy significantly. Unfortunately, FP needs to be executed within this NC FIFO analysis very often and each time it creates an exponentially growing set of alternative networks with prolongations. FP therefore does not scale and has been out of reach for the exhaustive analysis of large networks. We introduce DeepFP, an approach to make FP scale by predicting prolongations using machine learning. In our evaluation, we show that DeepFP can improve results in FIFO networks considerably. Compared to the aforementioned LUDB analysis, DeepFP reduces delay bounds by$12.1 \;\%$on average at negligible additional computational cost.

Precise indoor positioning systems (IPSs) are key to perform a set of tasks more efficiently during aircraft production, operation and maintenance. For instance, IPSs can overcome the tedious task of configuring (wireless) sensor nodes in an aircraft cabin. Although various solutions based on technologies of established consumer goods, e.g., Bluetooth or WiFi, have been proposed and tested, the published accuracy results fail to make these technologies relevant for many use cases. This stems from the challenging environments for positioning, especially in aircraft cabins, which is mainly due to the geometries, many obstacles, and highly reflective materials. To address these issues, we propose to evaluate in this work an Ultra-Wideband (UWB)-based IPS via a measurement campaign performed in a real aircraft cabin. We first illustrate the difficulties that an IPS faces in an aircraft cabin, by studying the signal propagation effects which were measured. We then investigate the ranging and localization accuracies of our IPS. Finally, we also introduce various methods based on machine learning (ML) for correcting the ranging measurements and demonstrate that we are able to localize a node with respect to an aircraft seat with a measured likelihood of 97 %.

Emerging Remote Piloting (RP) operations of electrified Unmanned Aerial Vehicles (UAVs) demand low-latency and high-quality video delivery to conduct safe operations in the low-altitude airspace. Although cellular networks are one of the prominent candidates to provide connectivity for such operations, their ground-centric nature limits their capabilities in achieving seamless and reliable aerial connectivity. In this paper, we study the feasibility of supporting RP operations with low latency and high-quality video delivery over commercial cellular networks. By setting up an adaptive bitrate video transmission pipeline with the Google Congestion Control (GCC) and Self-Clocked Rate Adaptation for Multimedia (SCReAM) Congestion Control (CC) algorithms, we analyze the video delivery performance for the RP application requirements and compare the performance of GCC and SCReAM against constant bitrate video delivery. Our results show that low-latency video delivery with < 300 ms playback latency between full-HD and 4K resolution can be maintained up to about 95% of the time in the air. While static bitrate video delivery outperforms adaptive streaming in urban location with abundant link capacity, the latter becomes advantageous in rural locations, where the link capacity is affected by fluctuations. Although the study's findings highlight the capabilities of cellular networks in delivering low-latency video for a safety-critical aerial service, we also discuss the potential improvements and future research challenges for enabling safe operations and meeting the service requirements using cellular networks. We release our collected traces and the video transmission pipeline as open-source to facilitate research in this field.

With the advent of standards for deterministic network behavior, synthesizing network designs under delay constraints becomes the natural next task to tackle. Network Calculus (NC) has become a key method for validating industrial networks, as it computes formally verified end-to-end delay bounds. However, analyses from the NC framework were thus far designed to bound one flow’s delay at a time. Attempts to use classical analyses for derivation of a network configuration revealed this approach to be poorly fitted for practical use cases. Take finding a delay-optimal routing configuration: One model for each routing alternative had to be created, then each flow delay had to be bounded, then the bounds were compared to the given constraints. To overcome this three-step procedure, we introduce Differential Network Calculus. We extend NC to allow for differentiation of delay bounds w.r.t. to a wide range of network parameters – such as flow routes. This opens up NC to a class of efficient nonlinear optimization techniques taking advantage of the delay bound computation’s gradient. Our numerical evaluation on the routing problem shows that our novel method can synthesize flow path in a matter of seconds, outperforming existing methods by multiple orders of magnitude.

Network support for Unmanned Aerial Vehicles (UAVs) is raising an interest among researchers due to the strong potential applications. However, current knowledge on UAV data traffic is mainly based on conceptual studies and does not provide an in-depth insight on the data traffic properties. To close this gap, we present a measurement-based study analyzing in detail the Control and Non-payload Communication (CNPC) traffic produced by three different UAVs when communicating with their remote controller over 802.11 protocol. We analyze the traffic in terms of data rate, inter-packet interval and packet length distributions, and identify their main influencing factors. The data traffic appears neither deterministic nor periodic but bursty, with a tendency towards Poisson traffic. We further create an understanding on how the traffic of the investigated UAVs are internally generated and propose a model to analytically capture their traffic processes, which provides an explanation for the observed behavior. We implemented a publicly available UAV traffic generator "AVIATOR" based on the proposed traffic model and verified the model by comparing the simulated traces with the experimental results.

Processing packets in batches is a common technique in high-speed software routers to improve routing efficiency and increase throughput. With the growing popularity of novel paradigms such as Network Function Virtualization, advocating for the replacement of hardware-based networking modules towards software-based network functions deployed on commodity servers, we observe that batching techniques have been successfully implemented to reduce the HW/SW performance gap. As batch creation and management is at the very core of high-speed packet processors, it provides a significant impact to the overall packet processing capabilities of the system, affecting latency, throughput, CPU utilization and power consumption. It is commonly accepted to adopt a fixed maximum batching size (usually in the range between 32 and 512) to optimize for the worst case scenario (i.e. minimum-size packets at full bandwidth capacity). Such approach may result in a loss of efficiency despite a 100% utilization of the CPU. In this work we explore the possibilities of enhancing the runtime batch creation in VPP, a popular software router based on the Intel DPDK framework. Instead of relying on the automatic batch creation, we apply machine learning techniques to optimize the batching size for lower CPU-time and higher power efficiency in average scenarios, while maintaining its high performance in the worst case.

Network calculus offers the means to compute worst-case traversal times based on interpreting a system as a queueing network. A major strength of network calculus is its strict separation of modeling and analysis frameworks. That is, a model is purely descriptive and can be put into multiple different analyses to derive a data flow's worst-case traversal time bound. One of the recent results in this category is the so-called flow prolongation. Flow prolongation actively manipulates the internal model of the analysis by virtually extending the path of flows, i.e., by deliberately creating a more pessimistic setting of resource contention between flows. It was shown that flow prolongation can theoretically decrease worst-case traversal time bounds under certain assumptions. Yet, due to its exhaustive search, it was also shown that flow prolongation does not scale and it might not even have an impact in larger queueing networks. In this paper we introduce DeepFP, an approach to make the analysis scale by predicting flow prolongations using a graph neural network. In our evaluation, we show that DeepFP can improve results in networks of FIFO queues considerably, where the delay bound can be reduced by 13.7% in large FIFO networks at negligible additional cost on the execution time of the analysis.

The network calculus (NC) analysis takes a simple model consisting of a network of schedulers and data flows crossing them. A number of analysis "building blocks" can then be applied to capture the model without imposing pessimistic assumptions like self-contention on tandems of servers. Yet, adding pessimism cannot always be avoided. To compute the best bound on a single flow’s end-to-end delay thus boils down to finding the least pessimistic contention models for all tandems of schedulers in the network – and an exhaustive search can easily become a very resource intensive task. The literature proposes a promising solution to this dilemma: a heuristic making use of machine learning (ML) predictions inside the NC analysis.While results of this work were promising in terms of delay bound quality and computational effort, there is little to no insight on when a prediction is made or if the trained algorithm can achieve similarly striking results in networks vastly differing from its training data. In this paper, we address these pending questions. We evaluate the influence of the training data and its features on accuracy, impact and scalability. Additionally, we contribute an extension of the method by predicting the best n contention model alternatives in order to achieve increased robustness for its application outside the training data. Our numerical evaluation shows that good accuracy can still be achieved on large networks although we restrict the training to networks that are two orders of magnitude smaller.

P4 introduces a standardized, universal way for data plane programming. Secure and resilient communication typically involves the processing of payload data and specialized cryptographic hash functions. We observe that current P4 targets lack the support for both. Therefore, applications and protocols, which require message authentication codes or hashing structures that are resilient against attacks such as denial-of-service, cannot be implemented. To enable authentication and resilience, we make the case for extending P4 targets with cryptographic hash functions. We propose an extension of the P4 Portable Switch Architecture for cryptographic hashes and discuss our prototype implementations for three different P4 target platforms: CPU, NPU, and FPGA. To assess the practical applicability, we conduct a performance evaluation and analyze the resource consumption. Our prototype implementations show that cryptographic hashing can be integrated efficiently. We cannot identify a single hash function delivering satisfying performance on all investigated platforms. Therefore, we recommend a set of hash functions to optimize target-specific performance.

Network calculus computes end-to-end delay bounds for individual data flows in networks of aggregate schedulers. It searches for the best model bounding resource contention between these flows at each scheduler. Analyzing networks, this leads to complex dependency structures and finding the tightest delay bounds becomes a resource intensive task. The exhaustive search for the best combination of contention models is known as Tandem Matching Analysis (TMA). The challenge TMA overcomes is that a contention model in one location of the network can have huge impact on one in another location. These locations can, however, be many analysis steps apart from each other. TMA can derive delay bounds with high degree of tightness but needs several hours of computations to do so. We avoid the effort of exhaustive search altogether by predicting the best contention models for each location in the network. For effective predictions, our main contribution in this paper is a novel framework combining graph-based deep learning and Network Calculus (NC) models. The framework learns from NC, predicts best NC models and feeds them back to NC. Deriving a first heuristic from this framework, called DeepTMA, we achieve provably valid bounds that are very competitive with TMA. We observe a maximum relative error below 6%, while execution times remain nearly constant and outperform TMA in moderately sized networks by several orders of magnitude.

With the increasing complexity of communication networks and the resulting threat of disruptions of mission critical services due to manual misconfiguration, automated verification is becoming a key element in today's network operation. In particular, it has recently been shown that a polynomial-time, automated verification of the policy-compliance of network configurations is possible for the important class of MPLS networks, even under failures. However, this approach, while providing polynomial runtimes, is still fairly slow in practice and only allows to detect but not fix configurations. This paper proposes a novel approach to speed up the analysis of network properties as well as to suggest configuration changes in case a network property is not satisfied. More specifically, our solution, DeepMPLS, allows to predict if a network property is satisfiable, and if not, aims to present a counter example. We also show that DeepMPLS may be used to propose new prefix-rewriting rules in the MPLS configuration in order to make it satisfiable. DeepMPLS can hence be used for fast predictions, before more rigorous analyses are performed. DeepMPLS is based on a new extension of graph-based neural networks. Our prototype implementation, using Tensorflow, achieves low execution times and high accuracies in real-world network topologies.

One of the major tasks when deploying real-time Ethernet networks is their configuration to achieve real-time behavior. In this paper we present an approach for a self-configuring plug-n-play network that automatically sets up devices and offers hard real-time guarantees to such devices. A new architecture and a protocol based on OpenFlow are proposed to achieve a system that can react to failures of switches and links by trying to repair such failures on a network level and restoring a full redundancy level while maintaining hard real-time guarantees. Implementation details are explained and the architecture is evaluated against randomly generated topologies. It is shown, that the solution can achieve a seamless failure recovery in link failures without any complexity at the end-device side.

Due to special requirements avionic networking devices are typically quite expensive. One way to reduce costs is to make use of commercial off the shelf devices and configure them in a way that gives similar performance. In this paper we evaluate the use of OpenFlow in the avionics environment in terms of performance and configuration. The main feature of OpenFlow is fine-grained access to the switch's forwarding plane. While it was primarily designed to offer high configurability and reduction of cost through harmonization of interfaces, in newer versions OpenFlow added support for traffic policing. In OpenFlow this is realized with meters that allow for quality of service enforcement on a hardware level as well as an arbitrary mapping of meters to flows. This paper shows how to make use of OpenFlow's meter commands to achieve deterministic behavior and discusses its advantages and shortcomings. We then implement the proposed solution on a commercial off the shelf OpenFlow switch and compare the switching performance to a state of the art avionics switch used in current aircraft.

We present in this paper a stochastic flow-level network model for the performance evaluation of IP networks with multiple bottlenecks supporting short-lived and long-lived TCP flows. Flow-level network models are efficient at estimating the mean bandwidth of TCP flows in various topologies, but they are generally limited to the study of infinite flows. This paper extends such models in order to evaluate short-lived flows alternating between idle and active periods where data of random size is transferred. We study the interaction between multiple flows and derive mean bandwidths, durations of file transfer or average number of active flows. We first study a single bottleneck, and then extend our analysis to networks with multiple bottlenecks as well as the effect of slow-start. We apply our results to various networks and assess the accuracy of our approach by comparing our analytical results with results of the discrete event simulator ns-2.

Possible evolution of Avionics Full-Duplex Switched Ethernet (AFDX) are currently investigated. Such network should support existing time-constrained avionics flows, as well as new best effort traffic, requiring scheduling algorithms to meet guarantees on avionics flows. Following the current trend of the automotive industry to move toward Audio/Video Bridging (AVB), this paper evaluates an application of the scheduling scheme described in the emerging IEEE 802.1 AVB standard in an avionics switched Ethernet scenario. Results are compared to other common scheduling strategies such as strict priority queuing, and different fair scheduling variants.