Since their deployment and commercialization, 5th Generation (5G) mobile systems have been extensively analyzed to quantify the Quality of Service and Experience (QoS/QoE) achievable by heterogeneous services. Real-time interactive services, i.e., applications within the scope of Ultra-Reliable Low Latency Communication (URLLC) and at the intersection of URLLC and enhanced Mobile Broadband (eMBB), are, however, often tested using simplistic methodologies that do not provide accurate assessments. In this paper, we extend our previous work on the empirical characterization of mobile networks by presenting a comprehensive analysis of a methodology standardized by the International Telecommunication Union Telecommunication Standardization Sector (ITU-T). This methodology is designed for systematic and reproducible QoS/QoE evaluations of real-time interactive services. We validate it through dedicated measurements (for which we open-source the corresponding dataset along with this paper) in the Karlstad University testbed, i.e., a private network supporting 5G connectivity modes and features beyond those available in current public networks in Sweden. Our results, spanning across services, mobility scenarios, connectivity modes, and servers, provide key insights into the intricate dependencies between QoS/QoE, environmental conditions, and system configurations, ultimately serving as a foundation for designing high-performing beyond-5G mobile systems.

Ensuring reliable real-time video streaming over heterogeneous networks particularly those combining 5G and satellite (Satcom) links poses significant challenges due to dynamic bandwidth, latency, and loss characteristics. This paper conducts an in-depth performance analysis of existing Multipath QUIC (MP-QUIC) schedulers in a hybrid 5G–Satcom environment, using a controlled emulation setup based on real network measurements from two operator datasets. We evaluate scheduler behavior across key Quality of Experience (QoE) metrics, including Peak Signal-to-Noise Ratio (PSNR), frame rate (FPS), and bitrate stability. Performance variability is captured using cumulative distribution functions (CDFs) across 50 experimental runs to ensure reproducibility. Our results reveal that while static rule-based schedulers (e.g., Round Robin, MinRTT) offer stable but conservative performance, they underutilize high-throughput paths in dynamic environments. Conversely, adaptive schedulers (e.g., Peekaboo, DEAR) exploit link diversity more aggressively, improving throughput under dynamic conditions. These observations illustrate critical trade-offs between path utilization and QoE stability. The study provides actionable insights for future multipath scheduling frameworks and underscores the potential of incorporating context-aware or learning-based decision mechanisms to improve video delivery resilience in complex, hybrid network scenarios. 

The increasing reliance on and integration of heterogeneous access networks such as 5G, Beyond 5G (B5G), and satellite communication (Satcom) systems have not been adequately leveraged in existing multipath transport schedulers. Although protocols like Multipath TCP (MPTCP) and Multipath QUIC (MPQUIC) enable bandwidth aggregation across multiple paths, current schedulers largely rely on reactive decisions and struggle to cope with rapidly varying hybrid network conditions involving bandwidth fluctuations, intermittent outages, and heterogeneous link characteristics. This work presents PRISM (Proximal policy optimization-based Reinforcement learning for Intelligent Scheduling of Multipaths), a deep reinforcement learning (DRL)-based scheduler designed for MPQUIC in hybrid 5G and Satcom environments. PRISM follows a hybrid-adaptive approach by integrating LSTM-enhanced actor–critic learning with an adaptive exploration strategy, allowing the scheduler to leverage temporal network behavior and make proactive scheduling decisions. The design remains lightweight, enabling efficient operation with low CPU and memory overhead, which is critical for real-time deployment. We evaluate PRISM under diverse and dynamically changing network conditions, including bandwidth variability and link disruptions, using real-world traces from Lumos5G and Starlink Satcom datasets. The evaluation compares PRISM against state-of-the-art learning-based schedulers (Peekaboo and DEAR) as well as widely used rule-based schedulers (RR, ECF, BLEST, and MinRTT). Experimental results show that PRISM consistently achieves superior performance, providing improvements of up to 25.64%–31.25% over other learning-based schedulers and substantially higher gains over rule-based approaches across heterogeneous network scenarios. 

The mobile telecommunication industry is undergoing continuous evolution to cope with ever increasing service requirements and expectations of end users. This has recently led to the rise of Mobile Network Aggregators (MNAs), a new type of global virtual operators that deliver mobile communication services by utilizing multiple Mobile Network Operators (MNOs), dynamically connecting to the one that best meets their customers' needs based on location and time. MNAs can then offer optimized global coverage by connecting to local MNOs that have limited (e.g., national) geographic service. In this paper, we provide a first in-depth analysis of the operations of three major MNAs: Google Fi, Twilio, and Truphone. We conduct performance measurements across these MNAs for critical applications spanning DNS, Web browsing, and video streaming, and compare their performance against that of a traditional MNO from two very diverse geographical locations, US and Spain. We find that MNAs may introduce some delay compared to local MNOs in the region where the user is roaming, yet they offer significant performance improvements over the traditional MNOs roaming model, such as home-routed roaming. To fully assess the potential benefits of the MNA model, we also carry out emulation studies assessing the potential performance gains that MNAs could achieve by deploying both control and user plane functions of open-source 5G implementations across different Amazon Web Services locations.

Virtual Network Functions (VNFs) are essential components in modern networking that decouple network functions from dedicated hardware, thereby enhancing flexibility, scalability, and cost-efficiency. The inherently dynamic nature and fluctuating loads of networks, combined with evolving network demands, often lead to the over-allocation or underprovisioning of VNF resources, posing a significant challenge in optimal VNF resource usage. This issue becomes even more challenging when VNFs are connected in a specific sequence to fulfill a functional need-Service Function Chain (SFC). A promising solution to this challenge is proactive resource management-predicting resource demands of VNFs in advance. In this paper, we present a performance metrics prediction framework designed to anticipate multi-step resource demands for VNFs. The framework employs machine learning models, including RNN, LSTM, GRU, and Transformer models, within a robust meta-learner to capture complex usage patterns, accounting for both short-term and long-term dependencies in VNF resource consumption. This approach allows for precise prediction of critical key performance indicators (KPIs), including CPU usage, memory usage, processing latency, and traffic load for each VNF within an SFC. The evaluation results show that the proposed framework achieves a 75% reduction in mean absolute error (MAE) compared to the Transformer model and over 84% compared to RNN, LSTM, and GRU models. These results demonstrate the framework's substantial improvement over state-of-the-art approaches. 

The Narrowband Internet of Things (NB-IoT) technology has an important role in the mobile cellular ecosystem, enabling massive Machine Type Communication (mMTC) services. NB-IoT propagation was preliminarily analyzed via a measurement campaign carried out in 2020 in the city of Oslo, Norway. This investigation resulted in Oslo-2020, the first NB-IoT-specific Alpha-Beta-Gamma (ABG) path loss (PL) model, which showed higher prediction accuracy compared to models developed for different technologies but often used for NB-IoT. In this paper, to further investigate NB-IoT PL in urban scenarios, we analyze new measurement campaigns performed in 2020-2021 and 2023 in the city of Rome, Italy. First, we use the 2020-2021 measurements to derive Rome-2021, a new NB-IoT-specific ABG PL model. We show that Rome-2021 preserves the statistical properties of Oslo-2020 (e.g., the Gaussianity of the PL exponent distribution across base stations), although the moments of the distributions are different due to city-specific environmental characteristics. We also use new data on signal losses due to outdoor-to-indoor propagation to refine the analysis of this scenario. Finally, we propose a methodology to combine Oslo-2020 and Rome-2021 into a more general model. Our methodology uses so-called Mixture Distributions (MDs), thus leveraging the shared statistical properties between Oslo-2020 and Rome-2021. By using the 2023 measurements, we show that our MD-based approach estimates PL model parameters with higher accuracy compared to Oslo-2020 and Rome-2021 models used separately, thus providing an effective solution for predicting NB-IoT urban PL in lack of site-specific measurements and information. 

Multi-Access Edge Computing (MEC) is a promising solution for providing the computational resources and low latency required by vehicular services, such as autonomous driving. It enables cars to offload computationally intensive tasks to nearby servers. Effective offloading involves determining when to offload tasks, selecting the appropriate MEC site, and efficiently allocating resources to ensure optimal performance. While car mobility poses significant challenges to guaranteeing reliable task completion, today we lack energy-efficient solutions to solve this problem, especially when considering real-world car mobility traces. In this paper, we begin by examining the mobility patterns of cars using data obtained from a leading mobile network operator in Europe. Based on the insights from this analysis, we design an optimization problem for task computation/offloading, considering both static and mobility scenarios. Our objective is to minimize the total energy consumption - both at the cars and the MEC nodes - while satisfying the latency requirements of various tasks. We evaluate our solution, based on multi-agent reinforcement learning, both in simulations as well as in a realistic setup that relies on datasets from the operator. Our solution shows a significant reduction of user dissatisfaction and task interruptions in both static and mobile scenarios, while achieving energy savings of 47% (static) and 14% (mobile) compared to state-of-the-art schemes. 

The promises of multipath transport are to aggregate bandwidth, improve resource utilisation and enhance reliability. In this article, we demonstrate that the way multipath coupled congestion control is defined today leads to a suboptimal resource utilisation when network paths are disjoint, i.e., they do not share a bottleneck link. With growing interest in standardising Multipath QUIC (MPQUIC), we have implemented the practical experiments, we evaluate MPQUIC-SBD in the context of video streaming with various Adaptive Bitrate (ABR) algorithms, addressing both ABR classes of rule- and learning-based solutions. We demonstrate that MPQUIC-SBD accurately detects shared bottlenecks over 90% of the time, depending on the ABR algorithm, as the size of the video segments increases. In non-shared bottleneck scenarios, when MPQUIC-SBD detects that its QUIC subflows do not share the same network resources, it decouples their congestion windows accordingly, enabling video throughput gains of up to 37% compared to MPQUIC. These gains translate directly into improved video quality metrics, including higher bitrate, better resolution and reduced buffering, resulting in an enhanced quality of experience for users.

Mobile Augmented Reality (MAR) applications pose unique challenges due to computation intensity, constrained device resources, and high interactive rendering requirements. The emergence of 5 G and edge computing offers opportunities to offload computation to the edge and cloud, indirectly enhancing the computing capability and usage duration of MAR devices. However, existing general task offloading and multipath transmission techniques do not address the challenges in offloading path selection with multiple edges, dynamic resource competition awareness, and spatial computation with strong task dependencies. This paper contributes FPSelector, a flexible path selector for MAR offloading. We present a two-tier MAR-specific offloading scheme with multiple edge nodes. In offloading decisions, we design a reinforcement learning model to generate the selection policy for each packet of an AR data stream. This model incorporates an action masking mechanism, a comprehensive reward function, and state features complemented by a resource prediction module, making FPSelector aware of dynamic heterogeneous environments. Moreover, we propose an online learning strategy to facilitate real-time selection. To validate its efficacy, we compare FPSelector’s performance against leading schedulers under various scenarios, demonstrating a notable reduction of 9.9% and 9.6% in overall completion time for 4 K and 8 K video-based MAR applications compared to its closest competitor. 

The advent of 5G and beyond 5G (B5G) systems has led to a significant rise in bandwidth-intensive applications within the Internet of Things (IoT), particularly in vehicular IoT (V-IoT). Effective solutions like Multipath TCP (MPTCP) and Multipath QUIC (MPQUIC) have emerged to address the escalating bandwidth demands of connected vehicles. However, challenges persist for multipath schedulers in efficiently adapting to diverse network conditions typically found in vehicular environments. In this paper, we introduce two novel variants of DEAR (Deep reinforcement learning Empowered Actor-critic scheduleR), namely, DEAR-MAC (Multiple Alternative Critics) and DEAR-CAP (Critic Associated per Path). The proposed DRL-based schedulers are tailored for multipath QUIC in 5G/B5G environments, enhancing decision-making in dynamic network scenarios often encountered by V-IoT devices. Through extensive experimentation across various network setups, including those with fluctuating bandwidth and network outages, and utilizing real-world network traces from the Lumos5G dataset, we conduct a comparative analysis against state-of-the-art learning-based schedulers like Peekaboo and rule-based schedulers like RR, ECF, BLEST, and minRTT. Our experiments show that the proposed DEAR-MAC and DEAR-CAP schedulers outperformed Peekaboo by 38.88% to 48.11%, respectively, in different heterogeneous network conditions, and the gains are much higher when compared to other rule-based schedulers. These advancements are particularly beneficial for vehicular IoT applications, ensuring more reliable and efficient data transmission, even in challenging network environments for applications such as real-time navigation, remote diagnostics, and vehicle-to-vehicle communication.