Search

Optical point-to-multipoint (P2MP) connectivity based on digital subcarrier multiplexing (DSCM) has been shown as a solution for the metro-access segment that is able to reduce capital and operational costs and support the capacity and high dynamicity needs of future 6G services. To achieve maximum performance, activation and deactivation of subcarriers must be done near-real-time to provide just the capacity needed to support the input traffic. In this paper, we investigate the applicability of various approaches capable of supporting the near-real-time operation requirement. Starting from the centralized approach that can be carried out on the centralized software-defined networking (SDN) controller, we also explore distributed approaches that might relieve the SDN controller from near-real-time operation. In particular, we explore the performance of deploying a multiagent system (MAS), where intelligent agents run on top of the nodes in the P2MP tree and communicate among them. Illustrative results show that the distributed approaches can achieve a performance close to that of the centralized one, while reducing communication needs. Results also show the importance of traffic/capacity prediction to anticipate the activation of subcarriers., The research leading to these results has received funding from the Smart Networks and Services Joint Undertaking under the European Union's Horizon Europe research and innovation programme under Grant Agreement No. 101096120 (SEASON), the MICINN IBON (PID2020-114135RB-I00) project and from the ICREA Institution., Peer Reviewed, Postprint (author's final draft)

The increase in the demand for eXtended Reality (XR)/Virtual Reality (VR) services in the recent years, poses a great challenge for Wi-Fi networks to maintain the strict latency requirements. In VR over Wi-Fi, latency is a significant issue. In fact, VR users expect instantaneous responses to their interactions, and any noticeable delay can disrupt user experience. Such disruptions can cause motion sickness, and users might end up quitting the service. Differentiating interactive VR traffic from Non-VR traffic within a Wi-Fi network can aim to decrease latency for VR users and improve Wi-Fi Quality of Service (QoS) with giving priority to VR users in the access point (AP) and efficiently handle VR traffic. In this paper, we propose a machine learning-based approach for identifying interactive VR traffic in a Cloud-Edge VR scenario. The correlation between downlink and uplink is crucial in our study. First, we extract features from single-user traffic characteristics and then, we compare six common classification techniques (i.e., Logistic Regression, Support Vector Machines, k-Nearest Neighbors, Decision Trees, Random Forest, and Naive Bayes). For each classifier, a process of hyperparameter tuning and feature selection, namely permutation importance is applied. The model created is evaluated using datasets generated by different VR applications, including both single and multi-user cases. Then, a Wi-Fi network simulator is used to analyze the VR traffic identification and prioritization QoS improvements. Our simulation results show that we successfully reduce VR traffic delays by a factor of 4.2x compared to scenarios without prioritization, while incurring only a 2.3x increase in delay for background (BG) traffic related to Non-VR services., This work is partially funded by Wi-XR PID2021-123995NB-I00 (MCIU/AEI/FEDER,UE), MAX-R ( 101070072 ) EU, SGR 00955-2021 AGAUR, and by MCIN/AEI under the Maria de Maeztu Units of Excellence Programme ( CEX2021-001195-M ). This paper has also been partially funded by the Spanish Ministry of Science and Innovation MCIN/AEI/10.13039/501100011033 under ARTIST project (ref. PID2020- 115104RB-I00 )., Peer Reviewed, Postprint (author's final draft)

This paper presents a study on connection-blocking prediction in Elastic Optical Networks (EONs) using Convolutional Neural Networks (CNNs). In EONs, connections are established and torn down dynamically to fulfill the instantaneous requirements of the users. The dynamic allocation of the connections may cause spectrum fragmentation and lead to network performance degradation as connection blocking increases. Predicting potential blocking situations can be helpful during EON operations. For example, this prediction could be used in real networks to trigger proper spectrum defragmentation mechanisms at suitable moments, thereby enhancing network performance. Extensive simulations over the well-known NSFNET (National Science Foundation Network) backbone network topology were run by generating realistic traffic patterns. The obtained results are later used to train the developed machine learning models, which allow the prediction of connection-blocking events. Resource use was continuously monitored and recorded during the process. Two different Convolutional Neural Network models, a 1D CNN (One-Dimensional Convolutional Neural Network) and 2D CNN (Two-Dimensional Convolutional Neural Network), are proposed as the predicting methods, and their behavior is compared to other conventional models based on an SVM (Support Vector Machine) and KNN (K Nearest Neighbors). The results obtained show that the proposed 2D CNN predicts blocking with the best accuracy (92.17%), followed by the SVM, the proposed 1D CNN, and KNN. Results suggest that 2D CNN can be helpful in blocking prediction and might contribute to increasing the efficiency of future EON networks., This research was partially funded by the Catalonia Government through the Agency for Management of University and Research Grants, grant SGR-2021-00598., Peer Reviewed, Postprint (published version)

The continuous evolution of the communications networks in the race to support 6G services will introduce important operational challenges for network operators, especially in terms of efficient usage of resources, near real-time performance, etc. Such advanced 6G services will necessarily leverage on existing brownfield networks complemented with novel technologies and supportive tools. In this context, the concept of network should be understood in a broad sense, including both networking and computing capabilities enabling the connectivity of the customers to service functions or applications in a very dynamic way. Until recently, the main driver for networks design and operation has been the guarantee of throughput since bandwidth-based services have been dominant so far. With the advent of 5G, delay-based services have started to become relevant as well, as a consequence of the offerings of low latency and interactive services, such as tactile internet or augmented and virtual reality. Now, the advent of forthcoming beyond 5G (B5G) and 6G services is expected to introduce more stringent requirements motivating a new radical change on operations, adding new dimensions to service provision, such as delay-variation or precision-based services. All this complexity triggers the need of evolving the mode of operation by introducing smarter, faster, and educated decisions on the operational processes, pursuing an overall optimization on the usage of both networking and compute resources. In this context, the concept of Network Digital Twins (NDTs) is gaining traction as a perfect complement to the foreseen complex operation., This work is funded by the Spanish Ministry of Economic Affairs and Digital Transformation and the European Union - NextGenerationEU under projects OPTIMAIX_OaaS (Ref. TSI-063000-2021-34) and OPTIMAIX_NDT (Ref. TSI063000-2021-35)., Peer Reviewed, Postprint (author's final draft)

The rapid development of Wi-Fi technologies in recent years has caused a significant increase in the traffic usage. Hence, knowledge obtained from Wi-Fi network measurements can be helpful for a more efficient network management. In this paper, we propose a methodology to predict future values of some specific network metrics (e.g. traffic load, transmission failures, etc.). These predictions may be useful for improving the network performance. After data collection and preprocessing, the correlation between each target access point (AP) and its neighbouring APs is estimated. According to these correlations, either an only-temporal or a spatio-temporal based prediction is done. To evaluate the proposed methodology, real measurements are collected from 100 APs deployed in different university buildings for 3 months. Deep Learning (DL) methods (i.e. Convolutional Neural Network (CNN), Simple Recurrent Neural Network (SRNN), Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM), Transformer) are evaluated and compared for both temporal and spatio-temporal based predictions. Moreover, a hybrid prediction methodology is proposed using a spatial processing based on CNN and a temporal prediction based on RNN. The proposed hybrid methodology provides an improvement in the prediction accuracy at expenses of a slight increase in the Training Computational Time (TCT) and negligible in Prediction Computational Time (PCT)., This paper has been funded by the Spanish Ministry of Science and Innovation MCIN/AEI/10.13039/501100011033 under ARTIST project (ref. PID2020-115104RB-I00)., Peer Reviewed, Postprint (author's final draft)

The research and innovation related to fifth-generation (5G) networks that has been carried out in recent years has decided on the fundamentals of the smart slice in radio access networks (RANs), as well as the autonomous fixed network operation. One of the most challenging objectives of beyond 5G (B5G) and sixth-generation (6G) networks is the deployment of mechanisms that enable smart end-to-end (e2e) network operation, which is required for the achievement of the stringent service requirements of the envisioned use cases to be supported in the short term. Therefore, smart actions, such as dynamic capacity allocation, flexible functional split, and dynamic slice management need to be performed in tight coordination with the autonomous capacity management of the fixed transport network infrastructure. Otherwise, the benefits of smart slice operation (i.e., cost and energy savings while ensuring per-slice service requirements) might be cancelled due to uncoordinated autonomous fixed network operation. Notably, the transport network in charge of supporting slices from the user equipment (UE) to the core expands across access and metro fixed networks. The required coordination needs to be performed while keeping the privacy of the radio and fixed network domains, which is important in multi-tenant scenarios where both network segments are managed by different operators. In this paper, we propose a novel approach that explores the concept of context-aware network operation, where the slice control anticipates the aggregated and anonymized information of the expected slice operation that is sent to the fixed network orchestrator in an asynchronous way. The context is then used as the input for the artificial intelligence (AI)-based models used by the fixed network control for the predictive capacity management of optical connections in support of RAN slices. This context-aware network operation aims at enabling accurate and reliable autonomous fixed network operation un, This work is partially supported by the SNS Joint Undertaken under grant agreement No. 101096120 (SEASON), the MICINN IBON (PID2020-114135RB-I00) project, and from the ICREA Institution., Peer Reviewed, Postprint (published version)

Probabilistic constellation shaping (PCS) has emerged as an advanced modulation technique that provides a fine-grained software-defined trade-off between achievable spectral efficiency (SE) and transmission reach to deliver optimal channel capacity at any distance. This paper quantifies the network throughput benefits resulting from adopting PCS in future Flex-Grid over multicore fibre optical backbone networks, compared to using traditional uniform modulation formats. In particular, different inter-channel guard band width configurations are accounted in our study, aiming to set guidelines on the technological requirements imposed to network spectrum selective switches (SSS) to take full PCS advantage in future optical backbone networks., This publication is part of the Spanish I+D+i projects TRAINER-A and TRAINER-B (refs. PID2020-118011GB-C21 and PID2020-118011GB-C22), funded by MCIN/AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

This paper quantifies the potential reduction in the number of required transceivers that the use of probabilistic constellation shaping (PCS) can bring to Flex-Grid over multicore fiber (MCF) optical backbone networks, which can directly lead to a reduction in network capital expenditure (CAPEX). Results from two different backbone networks illustrate that PCS can significantly decrease the number of required transceivers compared to using traditional modulation formats, thanks to the finer granularity in spectral efficiency. The observed reduction becomes more prominent (up to 22.3%) as the traffic profile bit-rates and the network physical distances increase., This publication is part of the Spanish I+D+i projects TRAINER-A and TRAINER-B (refs. PID2020-118011GB-C21 and PID2020-118011GB-C22), funded by MCIN/AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

Considerable research and standardization efforts are being made to support time-sensitive traffic, e.g., generated by applications like Industry 4.0 and 5G fronthaul, on packet networks. This paper focuses on analyzing the impact of conveying time-sensitive traffic in operators’ networks when such traffic is mixed with best-effort traffic. In particular, extensions to a continuous G/G/1/k queue model are proposed to evaluate two different Ethernet technologies, synchronous and asynchronous, supporting time-sensitive flows in terms of their influence on the performance of best-effort traffic., This project has received funding from the Smart Networks and Services Joint Undertaking under the European Union's Horizon Europe research and innovation programme under Grant Agreement No. 101095890 (PREDICT-6G), from the Spanish MINECO and the European Union though the NextGeneration UNICO5G TIMING (TSI-063000-2021-145), and the AEI IBON (PID2020-114135RB-I00) projects and from ICREA., Peer Reviewed, Postprint (author's final draft)

The advent of 6G will revolutionize the way Radio Access Networks (RAN) will be operated. Expected massive small cell deployments and features, such as an adaptive functional splitting, are expected to change not only the volume but also the requirements of the traffic to be supported by the fixed transport network. This paper presents an insight into 6G RAN operation, focusing on how such operation will impact the autonomous operation of the fixed network. As concluding remarks of such analysis, key requirements and challenges of fixed network operation for B5G/6G scenarios are identified., The research leading to these results has received funding from the Smart Networks and Services Joint Undertaking under the European Union's Horizon Europe research and innovation programme under Grant Agreement No. 101096120 (SEASON), and the MICINN IBON (PID2020-114135RB-I00) projects and from the ICREA Institution., Peer Reviewed, Postprint (author's final draft)

In general, the availability of an accurate machine learning (ML) model plays a particularly important role in the development of new networking solutions and is one of the main drivers for the development of 5G and beyond networking. Although an option is to update the model once inaccurate data is detected, such approach requires high computational effort, specially once the data history is large. In this paper, we propose an approach that combines a traffic prediction model based on Long Short-Term Memory (LSTM) with an analysis module for dynamic connection capacity allocation. Once the model is generated, re-training can be triggered after inaccuracies are detected by the analysis module. Illustrative numerical results show the benefits from the proposed decision-based re-training approach to reduce the number of re-training rounds while maintaining model accuracy., This work was partially funded by MCIN/AEI/ 10.13039/501100011033 grant PID2021-126431OB-I00 (ANEMONE), and Generalitat de Catalunya grant 2021 SGR 00770. This work also was funded by the European Commission Horizon Europe SNS JU DESIRE6G project (G.A. 101096466), from the Spanish MINECO and the European Union though the NextGeneration UNICO5G TIMING (TSI-063000-2021-145), and the AEI IBON (PID2020-114135RB-I00) projects. The authors also acknowledge the support of the French Agence Nationale de la Recherche (ANR), under grant ANR-19-CE-25-0001-01 (ARTIC project)., Peer Reviewed, Postprint (author's final draft)

Network Services automation requires predictable Quality of Service (QoS) performance, measured in terms of throughput, delay and jitter, to allow making proactive decisions. QoS is typically guaranteed by overprovisioning capacity dedicated to the packet connection, which increases costs for customers and network operators, especially when the traffic generated by the users and/or the virtual functions highly varies over the time. This paper presents the PILOT methodology for modeling the performance of packet connections during commissioning testing in terms of throughput, delay and jitter. PILOT runs in a sandbox domain and constructs a scenario where an efficient traffic flow simulation environment, based on the CURSA-SQ model, is used to generate large amounts of data for Machine Learning (ML) model training and validation. The simulation scenario is tuned using real measurements of the connection obtained from a set of active probes., The research leading to these results has received funding from the Smart Networks and Services Joint Undertaking under the European Union's Horizon Europe research and innovation programme under Grant Agreement No. 101095890 (PREDICT6G), from the Spanish MINECO and the European Union though the NextGeneration UNICO5G TIMING (TSI-063000- 2021-145/148/149), and the AEI IBON (PID2020-114135RB-I00) projects and from ICREA., Peer Reviewed, Postprint (author's final draft)

The rise of traffic intensive services and applications is pushing the limits of conventional single band Wavelength Division Multiplexing (WDM) optical networks. As an answer to this challenge, new data plane technologies are being investigated. Multi-band optical networks have raised as a very interesting candidate due to the potential increased capacity they offer thanks to the exploitation of multiple bands of the optical spectrum. Considering the whole telecom ecosystem, multi-band optical networks will coexist with other technological segments (e.g., Radio Access Network (RAN)) with the aim of provisioning services across the end-to-end infrastructure. With the advent of 5G and beyond 5G (B5G) architectures, novel provisioning paradigms are taking preponderance, such as the case of network slicing, which represents a radical paradigm change with respect to legacy business and provisioning models. As such, proper solutions for supporting network slice provisioning and runtime maintenance at the data plane are required. With this in mind, in this paper we present a control and orchestration architecture for the configuration and maintenance of network slices in multi-band optical networks, in support of B5G end-to-end services. Indeed, quality assurance and maintenance at all levels is seen as a cornerstone in B5G architectures. Thus, proper mechanisms adapted to the nature of the underlying sliceable multi-band data plane are required to ensure the quality of deployed slices. In this regard, we also present a novel band-adaptive protection scheme which takes advantage of the properties of the multi-band data plane so as to enhance the robustness of slices against quality degradations. We showcase the provisioning and maintenance of multi-band optical network slices by means of an experimental demonstration in a real testbed deployed at our premises. In addition, we evaluate the performance of the proposed band-adaptive protection scheme for slice quality assura, This publication is part of the Spanish I+D+i project TRAINER-B (ref. PID2020-118011GB-C22), funded by MCIN/AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

Network monitoring solutions need to cope with increasing network traffic volumes, as a result, sketch-based monitoring methods have been extensively studied to trade accuracy for memory scalability and storage reduction. However, sketches are sensitive to skewness in network flow distributions due to hash collisions, and need complicated performance optimization to adapt to line-rate packet streams. We provide Jellyfish, an efficient sketch method that performs one-pass clustering over the network stream. One-pass clustering is realized by adapting the monitoring granularity from the whole network flow to fragments called subflows, which not only reduces the ingestion rate but also provides an efficient intermediate representation for the input to the sketch. Jellyfish provides the network-flow level query interface by reconstructing the network-flow level counters by merging subflow records from the same network flow. We provide probabilistic analysis of the expected accuracy of both existing sketch methods and Jellyfish. Real-world trace-driven experiments show that Jellyfish reduces the average estimation errors by up to six orders of magnitude for per-flow queries, by six orders of magnitude for entropy queries, and up to ten times for heavy-hitter queries., This work was supported in part by the National Natural Science Foundation of China (NSFC) under Grant 61972409; in part by Hong Kong Research Grants Council (RGC) under Grant TRS T41-603/20-R, Grant GRF-16213621, and Grant ITF ACCESS; in part by the Spanish I+D+i project TRAINER-A, funded by MCIN/AEI/10.13039/501100011033, under Grant PID2020-118011GB-C21; and in part by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed, Postprint (author's final draft)

Network models are an essential block of modern networks. For example, they are widely used in network planning and optimization. However, as networks increase in scale and complexity, some models present limitations, such as the assumption of Markovian traffic in queuing theory models, or the high computational cost of network simulators. Recent advances in machine learning, such as Graph Neural Networks (GNN), are enabling a new generation of network models that are data-driven and can learn complex non-linear behaviors. In this paper, we present RouteNet-Fermi, a custom GNN model that shares the same goals as Queuing Theory, while being considerably more accurate in the presence of realistic traffic models. The proposed model predicts accurately the delay, jitter, and packet loss of a network. We have tested RouteNet-Fermi in networks of increasing size (up to 300 nodes), including samples with mixed traffic profiles — e.g., with complex non-Markovian models — and arbitrary routing and queue scheduling configurations. Our experimental results show that RouteNet-Fermi achieves similar accuracy as computationally-expensive packet-level simulators and scales accurately to larger networks. Our model produces delay estimates with a mean relative error of 6.24% when applied to a test dataset of 1,000 samples, including network topologies one order of magnitude larger than those seen during training. Finally, we have also evaluated RouteNet-Fermi with measurements from a physical testbed and packet traces from a real-life network., This work was supported in part by the Spanish I+D+i Project TRAINER-A funded by the Ministry of Science and Innovation (MCIN)/State Investigation Agency (AEI)/10.13039/501100011033 under Grant PID2020-118011GB-C21, in part by the Catalan Institution for Research and Advanced Studies (ICREA), and in part by the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund., Peer Reviewed, Postprint (author's final draft)

The centralization of baseband (BB) functions in a radio access network (RAN) towards data processing centres is receiving increasing interest as it enables the exploitation of resource pooling and statistical multiplexing gains among multiple cells, facilitates the introduction of collaborative techniques for different functions (e.g., interference coordination), and more efficiently handles the complex requirements of advanced features of the fifth generation (5G) new radio (NR) physical layer, such as the use of massive multiple input multiple output (MIMO). However, deciding the functional split (i.e., which BB functions are kept close to the radio units and which BB functions are centralized) embraces a trade-off between the centralization benefits and the fronthaul costs for carrying data between distributed antennas and data processing centres. Substantial research efforts have been made in standardization fora, research projects and studies to resolve this trade-off, which becomes more complicated when the choice of functional splits is dynamically achieved depending on the current conditions in the RAN. This paper presents a comprehensive tutorial on the characterisation, modelling and assessment of functional splits in a flexible RAN to establish a solid basis for the future development of algorithmic solutions of dynamic functional split optimisation in 5G and beyond systems. First, the paper explores the functional split approaches considered by different industrial fora, analysing their equivalences and differences in terminology. Second, the paper presents a harmonized analysis of the different BB functions at the physical layer and associated algorithmic solutions presented in the literature, assessing both the computational complexity and the associated performance. Based on this analysis, the paper presents a model for assessing the computational requirements and fronthaul bandwidth requirements of different functional splits. Last, the model is use, This work has been partially funded by Huawei Technologies. Work by X. Gelabert and B. Klaiqi is partially funded by the European Union's Horizon Europe research and innovation programme (HORIZON-MSCA-2021-DN-0) under the Marie Skłodowska-Curie grant agreement No 101073265. Work by J. Perez-Romero and O. Sallent is also partially funded by the Smart Networks and Services Joint Undertaking (SNS JU) under the European Union’s Horizon Europe research and innovation programme under Grant Agreements No. 101096034 (VERGE project) and No. 101097083 (BeGREEN project) and by the Spanish Ministry of Science and Innovation MCIN/AEI/10.13039/501100011033 under ARTIST project (ref. PID2020-115104RB-I00). This last project has also funded the work by D. Campoy., Peer Reviewed, Postprint (author's final draft)

According to forecasts, around 10 billion connected devices in the world will exchange data over the Internet by2028 and the significant connections increase threatens to congest terrestrial networks. These devices will mainly beembedded, controlled by smart software and operated in the cloud, integrating the Internet of Things ecosystem. Addi-tionally, current terrestrial infrastructures do not meet novel IoT requirements: remote access areas, multiple networkdomains crossing and adaptable latency and bandwidth. To provide global and ubiquitous coverage, the 3GPP stan-dardizes in rel-17 the concept of Non-Terrestrial-Networks. The term includes, among others, Non-Terrestrial nodessuch as GEO, MEO and LEO satellites. Lately, LEO satellites have contributed to consolidate the New Space indus-try opening new business opportunities by providing services using low-cost CubeSats. Despite offering importantbenefits compared to GEO (e.g. lower latency and propagation loss), LEO still has challenges to face related to com-munication discontinuity (intermittent satellite-to-ground links). Global region coverage is achieved through satelliteconstellations; however, the main concern is the non-guarantee of simultaneous contact from UE to the Ground StationNetwork via the satellite. While 3GPP rel-17 specs have centered on the specification of satellites using transparentpayload, the use of regenerative payloads has not precluded and has a central role in the technology roadmap. Re-generative payloads envision to embark part of the NF within the satellite, giving multiple enhancements (e.g. lowerlatencies, Store-and-Forward operation, in-satellite local loop communications and edge computing). Onboard func-tions may include radio access functions, such as eNB/gNB, as well as core NF, which enables direct satellite responseto UE. This paper provides a performance analysis of a deployment of core NF in a regenerative payload with a Soft-ware Defined Radio based on FPGA technology, Peer Reviewed, Postprint (published version)

Current trends in networking propose the use of Machine Learning (ML) for a wide variety of network optimization tasks. As such, many efforts have been made to produce ML-based solutions for Traffic Engineering (TE), which is a fundamental problem in Internet Service Provider (ISP) networks. Nowadays, state-of-the-art TE optimizers rely on traditional optimization techniques, such as Local search, Constraint Programming, or Linear programming. In this paper, we present MAGNNETO, a distributed ML-based framework that leverages Multi-Agent Reinforcement Learning and Graph Neural Networks for distributed TE optimization. MAGNNETO deploys a set of agents across the network that learn and communicate in a distributed fashion via message exchanges between neighboring agents. Particularly, we apply this framework to optimize link weights in Open Shortest Path First (OSPF), with the goal of minimizing network congestion. In our evaluation, we compare MAGNNETO against several state-of-the-art TE optimizers in more than 75 topologies (up to 153 nodes and 354 links), including realistic traffic loads. Our experimental results show that, thanks to its distributed nature, MAGNNETO achieves comparable performance to state-of-the-art TE optimizers with significantly lower execution times. Moreover, our ML-based solution demonstrates a strong generalization capability to successfully operate in new networks unseen during training., This publication is part of the Spanish I+D+i project TRAINER-A (ref. PID2020-118011GBC21), funded by MCIN/AEI/10.13039/501100011033. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA), the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia, and the European Social Fund., Peer Reviewed, Postprint (author's final draft)

Wireless communications at the chip scale emerge as a interesting complement to traditional wire-based approaches thanks to their low latency, inherent broadcast nature, and capacity to bypass pin constraints. However, as current trends push towards massive and bandwidth-hungry processor architectures, there is a need for wireless chip-scale networks that exploit and share as many channels as possible. In this context, this work addresses the issue of channel sharing by exploring the design space of multi-channel Medium Access Control (MAC) protocols for chip-scale networks. Distinct channel assignment strategies for both random access and token passing are presented and evaluated under realistic traffic patterns. It is shown that, even with the improvements enabled by the multiple channels, both protocols maintain their intrinsic advantages and disadvantages., This work is supported by the European Commission under H2020 grant WiPLASH (GA 863337) and HE grant WINC (GA 101042080)., Peer Reviewed, Postprint (author's final draft)

In this work, we aim at investigating the applicability of machine learning (ML), namely, a linear regression (LR) model for estimation of flows latencies in packet-switched 5G xHaul networks. The analysis is performed in a network scenario in which the switches are connected using high-capacity fiber links., The work was supported by National Science Centre, Poland under Grant No. 2018/31/B/ST7/03456. Moreover, it was also supported by the Spanish I+D+i project TRAINER-A (ref. PID2020-118011GB-C21), funded by MCIN/AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

This work investigates the transparent bypassing capacity requirements of elastic backbone optical networks to determine the all-optical cross-connection capacity needed at network nodes. Topological parameters have been used to develop a random network generator, and the generated topologies are evaluated. Reference values are obtained and applied to well-known topologies, and extensive simulations are conducted to obtain network nodes' bypassing traffic under realistic traffic profiles. The study reveals that although bypassing traffic varies by topology, it never exceeds 9% of total network traffic per node degree. These findings can help properly dimension network nodes by determining the necessary quantity of transceivers and cross-connection capacity based on the network topology and expected traffic., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work has been partially funded by Spanish AEI IBON (PID2020-114135RB-I00)., Peer Reviewed, Postprint (published version)

This paper proposes a solution for dynamically allocating subcarriers in a point-to-multipoint connection in the upstream direction. Using Digital Subcarrier Multiplexing a single carrier can be divided into several subcarriers. An ILP was proposed to dynamically allocated subcarriers in various traffic scenarios allowing for an increased number of edge nodes serviced. Significant capital and energy cost reductions were shown in comparison to traditional methods., This project has received funding from the Smart Networks and Services Joint Undertaking under the European Union's Horizon Europe research and innovation programme under Grant Agreement No. 101096120 (SEASON), from the MICINN IBON (PID2020-114135RB-I00) project and from the ICREA Institution., Peer Reviewed, Postprint (author's final draft)

As the dynamicity of the traffic increases, the need for self-network operation becomes more evident. One of the solutions that might bring cost savings to network operators, is that of the dynamic capacity management of large packet flows, especially in the context of packet over optical networks. Machine Learning, and particularly Reinforcement Learning (RL), seem to be an enabler for autonomicity, as a result of its inherent capacity to learn from experience. In this tutorial, we introduce RL and review its application for autonomous capacity management of traffic flows., This project has received funding from the Smart Networks and Services Joint Undertaking under the European Union's Horizon Europe research and innovation programme under Grant Agreement No. 101096120 (SEASON), from the MICINN IBON (PID2020-114135RB-I00) project and from the ICREA Institution., Peer Reviewed, Postprint (author's final draft)

The recently created ITU-T Focus Group Network 2030 is leading network operators to identify the requirements and use cases that networks are expected to fulfill for the short, medium, and long term within the current decade. Essentially, network operators need to evolve their networks to meet strict performance requirements in several dimensions, including a large bandwidth to support foreseen beyond-5G (B5G) services, such as digital twins and volumetric video. To provide such a bandwidth requirement in a sustainable and scalable way, multi-band (MB) optical networks are expected to gradually extend legacy optical network capacity by exploiting bands beyond C+L. In this paper, we present a traffic analysis methodology to help network operators to compute expected traffic demand to be supported in their networks as a result of combining well-known mass market services with foreseen B5G service scenarios. Numerical results based on inputs and forecasts from major European network operators show that MB will be required at all network segments, including metro-aggregation, metro-core, and backbone, by the end of this decade., Agencia Estatal de Investigación [PID2020-114135RBI00 (IBON)]; Horizon 2020 Framework Programme [101016663 (B5G-OPEN)]; Institució Catalana de Recerca i Estudis Avançats., Peer Reviewed, Postprint (author's final draft)

† This manuscript is the extension of the conference paper: Iqbal, M.; Ruiz, M.; Costa, N.; Napoli, A.; Pedro, J.; Velasco, L. Dynamic and Efficient Point-to-Point and Point-to-Multipoint Communications by Slicing the Optical Constellation. In Proceedings of the Optical Fiber Communications Conference and Exhibition (OFC), San Diego, CA, USA, 6–10 March 2022., New 5 G and beyond services demand innovative solutions in optical transport to increase efficiency and flexibility and reduce capital (CAPEX) and operational (OPEX) expenditures to support heterogeneous and dynamic traffic. In this context, optical point-to-multipoint (P2MP) connectivity is seen as an alternative to provide connectivity to multiple sites from a single source, thus potentially both reducing CAPEX and OPEX. Digital subcarrier multiplexing (DSCM) has been shown as a feasible candidate for optical P2MP in view of its ability to generate multiple subcarriers (SC) in the frequency domain that can be used to serve several destinations. This paper proposes a different technology, named optical constellation slicing (OCS), that enables a source to communicate with multiple destinations by focusing on the time domain. OCS is described in detail and compared to DSCM by simulation, where the results show that both OCS and DSCM provide a good performance in terms of the bit error rate (BER) for access/metro applications. An exhaustive quantitative study is afterwards carried out to compare OCS and DSCM considering its support to dynamic packet layer P2P traffic only and mixed P2P and P2MP traffic; throughput, efficiency, and cost are used here as the metrics. As a baseline for comparison, the traditional optical P2P solution is also considered in this study. Numerical results show that OCS and DSCM provide a better efficiency and cost savings than traditional optical P2P connectivity. For P2P only traffic, OCS and DSCM are utmost 14.6% more efficient than the traditional lightpath solution, whereas for heterogeneous P2P + P2MP traffic, a 25% efficiency improvement is achieved, making OCS 12% more efficient than DSCM. Interestingly, the results show that for P2P only traffic, DSCM provides more savings of up to 12% than OCS, whereas for heterogeneous traffic, OCS can save up to 24.6% more than DSCM., This research was funded by European Commission through the H2020 MSCA REAL-NET (G.A. 813144) and the HORIZON ALLEGRO (G.A. 101092766) projects, by the AEI through the IBON project (PID2020-114135RB-I00), and by the ICREA institution., Peer Reviewed, Postprint (published version)

In the current contribution, we examine the feasibility of fully-energy-autonomous operation of reconfigurable intelligent surfaces (RIS) through wireless energy harvesting (EH) from incident information signals. Towards this, we first identify the main RIS energy-consuming components and present a suitable and accurate energy-consumption model that is based on the recently proposed integrated controller architecture and includes the energy consumption needed for channel estimation. Building on this model, we introduce a novel RIS architecture that enables EH through RIS unit-cell (UC) splitting. Subsequently, we introduce an EH policy, where a subset of the UCs is used for beamsteering, while the remaining UCs absorb energy. In particular, we formulate a subset allocation optimization problem that aims at maximizing the signal-to-noise ratio (SNR) at the receiver without violating the RIS’s energy consumption demands. As a problem solution, we present low-complexity heuristic algorithms. The presented numerical results reveal the feasibility of the proposed architecture and the efficiency of the presented algorithms with respect to both the optimal and very high-complexity brute-force approach and the one corresponding to random subset selection. Furthermore, the results reveal how important the placement of the RIS as close to the transmitter as possible is, for increasing the harvesting effectiveness., This work was supported by the Luxembourg National Research Fund (FNR) under the CORE project RISOTTI (ref. 14773976), the European Commission’s Horizon 2020 research and innovation programme (ARIADNE) under grant agreement No. 871464, and the Digital Futures center., Peer Reviewed, Postprint (published version)

As the number of wireless technologies (4G, 5G, 802.11ax and other) have been rapidly increasing, so have the number of wireless networks concurrently deployed on a given coverage area. As a result, the judicious selection of the network that maximizes the quality perceived by a user terminal has become a significantly relevant problem. Contextual Multi-Armed Bandits (CMAB) are viable models to approach the problem. While multiple CMAB algorithms have been designed, most of them are only suited for stationary environments. This work proposes a new set of network selection algorithms that relate the traffic type (used as contextual information) to the perceived quality of the available networks for non-stationary scenarios. Results show significantly improved performance when compared to non-adaptive approaches., This work has been funded through the project ROUTE56 (Agencia Estatal de Investigación, Ministerio de Ciencia e Innovación, PID2019-104945GB-I00/ AEI / 10.13039/501100011033) and the grants 22CO1/008248 and 2021 SGR 01033 (AGAUR, Generalitat de Catalunya)., Peer Reviewed, Postprint (author's final draft)

Network automation promises to reduce costs while guaranteeing the required performance; this is paramount when dealing with the forecasted highly dynamic traffic that will be generated by new 5G/6G applications. In optical networks, autonomous lightpath operation entails that the optical receiver can identify the configuration of a received optical signal without necessarily being configured from the network controller. This provides relief for the network controller from real-time operation, and it can simplify the operation in multi-domain scenarios, where an optical connection spans across more than one domain. Consequently, in this work, we propose a blind and low complex modulation format (MF) and symbol rate (SR) identification algorithm. The algorithm is based on studying the effects of decoding an optical signal with different MFs and SRs. Extensive MATLAB-based simulations have been carried out which consider a coherent wavelength division multiplexed system based on 32 and 64 quadrature amplitude modulated signals at up to 96 GBd, thus enabling bit rates of up to 800 Gb/s/channel. The results show remarkable identification accuracy in the presence of linear and nonlinear noise for a wide range of feasible configurations., This research was funded by European Commission through the H2020 MSCA REAL-NET (G.A. 813144) and the B5G-OPEN (G.A. 101016663) projects, by the AEI through the IBON project (PID2020-114135RB-I00), and by ICREA institution., Peer Reviewed, Postprint (published version)

This paper envisages a Beyond 5G (B5G) Radio Access Network (RAN) in which the relaying capabilities offered by user equipment (UE) are used as a way to improve the coverage and robustness of the network. The paper proposes and develops the functional framework for supporting the activation of the suitable relay UEs (RUEs) in coverage constrained scenarios. It is based on characterizing each potential RUE through a utility metric that measures the coverage enhancements brought to the network when the RUE is activated. To derive this metric for all the candidate RUEs, the framework considers the use of a Network Digital Twin that allows the offline analysis of different configurations in a fast and safe way. Using the proposed framework, a RUE activation algorithm is proposed and evaluated. The obtained results reflect that significant outage probability reductions can be obtained in the scenario under different traffic distributions thanks to the activation of the RUEs with the highest utility., This paper is part of ARTIST project (ref. PID2020- 115104RB-I00) funded by MCIN/AEI/10.13039/ 501100011033. The work is also funded by the Spanish Ministry of Economic Affairs and Digital Transformation and the European Union - NextGenerationEU under project OPTIMAIX_NDT (Ref. TSI-063000-2021-35)., Peer Reviewed, Postprint (author's final draft)

This paper envisions to augment the Radio Access Network (RAN) infrastructure in Beyond 5G(B5G) systems by exploiting relaying capabilities of user equipment (UE) as a way to improve the coverage, capacity and robustness. Despite the concept and enabling technologies have been in place for some time, their efficient realization requires the conception and development of new features in B5G systems. Among them, this paper focuses on the Relay UE (RUE) activation decision making, in charge of deciding where and when a UE is suitable to be activated to relay traffic from other UEs. Specifically, the paper analyses seven RUE activation strategies that differ on the criteria and the type of context information considered for this decision-making problem. The considered strategies are evaluated through system level simulations in a realistic urban scenario with the objective of assessing the value of each type of context information. Results reveal that the most efficient strategies from the perspective of outage probability reduction are those that account for the number of UEs that would be served by a RUE based on the experienced spectral efficiency., This paper is part of ARTIST project (ref. PID2020-115104RB-I00) funded by MCIN/AEI/10.13039/ 501100011033., Peer Reviewed, Postprint (author's final draft)

Network modeling is a critical component of Quality of Service (QoS) optimization. Current networks implement Service Level Agreements (SLA) by careful configuration of both routing and queue scheduling policies. However, existing modeling techniques are not able to produce accurate estimates of relevant SLA metrics, such as delay or jitter, in networks with complex QoS-aware queueing policies (e.g., strict priority, Weighted Fair Queueing, Deficit Round Robin). Recently, Graph Neural Networks (GNNs) have become a powerful tool to model networks since they are specifically designed to work with graph-structured data. In this paper, we propose a GNN-based network model able to understand the complex relationship between the queueing policy (scheduling algorithm and queue sizes), the network topology, the routing configuration, and the input traffic matrix. We call our model TwinNet, a Digital Twin that can accurately estimate relevant SLA metrics for network optimization. TwinNet can generalize to its input parameters, operating successfully in topologies, routing, and queueing configurations never seen during training. We evaluate TwinNet over a wide variety of scenarios with synthetic traffic and validate it with real traffic traces. Our results show that TwinNet can provide accurate estimates of end-to-end path delays in 106 unseen real-world topologies, under different queuing configurations with a Mean Absolute Percentage Error (MAPE) of 3.8%, as well as a MAPE of 6.3% error when evaluated with a real testbed. We also showcase the potential of the proposed model for SLA-driven network optimization and what-if analysis., This publication is part of the Spanish I+D+i project TRAINER-A (ref.PID2020-118011GB-C21), funded by MCIN/ AEI/, Spain10.13039/501100011033. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA), Spain and the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia, Spain and the European Social Fund., Peer Reviewed, Postprint (published version)

Network modeling is an essential tool for network planning and management. It allows network administrators to explore the performance of new protocols, mechanisms, or optimal configurations without the need for testing them in real production networks. Recently, Graph Neural Networks (GNNs) have emerged as a practical solution to produce network models that can learn and extract complex patterns from real data without making any assumptions. However, state-of-the-art GNN-based network models only work with traffic matrices, this is a very coarse and simplified representation of network traffic. Although this assumption has shown to work well in certain use-cases, it is a limiting factor because, in practice, networks operate with flows. In this paper, we present FlowDT a new DL-based solution designed to model computer networks at the fine-grained flow level. In our evaluation, we show how FlowDT can accurately predict relevant per-flow performance metrics with an error of 3.5%, FlowDT’s performance is also benchmarked against vanilla DL models as well as with Queuing Theory., This work has been supported by the Spanish Government through project TRAINER-A (PID2020-118011GB-C21) with FEDER contribution and the Catalan Institution for Research and Advanced Studies (ICREA)., Peer Reviewed, Postprint (author's final draft)

Optical Constellation Slicing is proposed to convey heterogenous traffic from a source to multiple destinations, while supporting dynamic capacity allocation. Illustrative numerical results reveal the potential of the proposed scheme, while providing significant cost reduction., The research leading to these results has received funding from the European Community through the MSCA REAL-NET project (G.A. 813144), by the AEI through the IBON project (PID2020-114135RB-I00), and by the ICREA institution., Peer Reviewed, Postprint (author's final draft)

Load imbalance is a long-standing source of inefficiency in high performance computing. The situation has only got worse as applications and systems increase in complexity, e.g., adaptive mesh refinement, DVFS, memory hierarchies, power and thermal management, and manufacturing processes. Load balancing is often implemented in the application, but it obscures application logic and may need extensive code refactoring. This paper presents an automated and transparent dynamic load balancing approach for MPI applications with OmpSs-2 tasks, which relieves applications from this burden. Only local and trivial changes are required to the application. Our approach exploits the ability of OmpSs-2@Cluster to offload tasks for execution on other nodes, and it reallocates compute resources among ranks using the Dynamic Load Balancing~(DLB) library. It employs LeWI to react to fine-grained load imbalances and DROM to address coarse-grained load imbalances by reserving cores on other nodes that can be reclaimed on demand. We use an expander graph to limit the amount of point-to-point communication and state. The results show 46% reduction in time-to-solution for micro-scale solid mechanics on 32 nodes and a 20% reduction beyond DLB for $n$-body on 16 nodes, when one node is running slow. A synthetic benchmark shows that performance is within 10% of optimal for an imbalance of up to 2.0 on 8 nodes. All software is released open source., This research has received funding from the European Union’s Horizon 2020/EuroHPC research and innovation programme under grant agreement No 955606 (DEEP-SEA) and 754337 (EuroEXA). It is supported by the Spanish State Research Agency - Ministry of Science and Innovation (contract PID2019-107255GB and Ramon y Cajal fellowship RYC2018-025628-I) and by the Generalitat de Catalunya (2017-SGR-1414)., Peer Reviewed, Postprint (author's final draft)

Emerging applications such as the metaverse, telesurgery or cloud computing require increasingly complex operational demands on networks (e.g., ultra-reliable low latency). Likewise, the ever-faster traffic dynamics will demand network control mechanisms that can operate at short timescales (e.g., sub-minute). In this context, Traffic Engineering (TE) is a key component to efficiently control network traffic according to some performance goals (e.g., minimize network congestion).This paper presents Routing By Backprop (RBB), a novel TE method based on Graph Neural Networks (GNN) and differentiable programming. Thanks to its internal GNN model, RBB builds an end-to-end differentiable function of the target TE problem (MinMaxLoad). This enables fast TE optimization via gradient descent. In our evaluation, we show the potential of RBB to optimize OSPF-based routing (˜25% of improvement with respect to default OSPF configurations). Moreover, we test the potential of RBB as an initializer of computationally-intensive TE solvers. The experimental results show promising prospects for accelerating this type of solvers and achieving efficient online TE optimization., This work was supported by the Polish Ministry of Science and Higher Education with the subvention funds of the Faculty of Computer Science, Electronics and Telecommunications of AGH University and by the PL-Grid Infrastructure. Also, this publication is part of the Spanish I+D+i project TRAINER-A (ref. PID2020-118011GB-C21), funded by MCIN/ AEI/10.13039/501100011033. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA) and the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund., Peer Reviewed, Postprint (author's final draft)

Fifth Generation (5G) networks will support countless new applications and new business models. One of the 5G paradigms is network slicing, which enables the integration of multiple logical networks each one tailored to the requirements of the different services that can be provided by both network operators and vertical industries. One of the services where 5G is expected to have a greatest impact is vehicular-to-everything (V2X) communications, which will have their stringent latency requirements now met. However, the mobility associated to vehicles can lead to cell overload compromising the required quality of service (QoS). To address this problem, in this paper we propose and evaluate the performance of three network overload alleviation techniques to control network congestion provoked by traffic jams using realistic vehicular traces in a network slicing environment. Firstly, we describe the architecture supporting V2X communications. Secondly, the network congestion control approaches are explained. Finally, after providing a complete description of the considered scenario, results will be detailed, showing that the network overload appearing during rush hour can be significantly reduced., This research was supported by the Spanish Centre for the Development of Industrial Technology (CDTI) and the Ministry of Economy, Industry and Competitiveness under grant/project CER-20191015 / Open, Virtualized Technology Demonstrators for Smart Networks (Open-VERSO)., Peer Reviewed, Postprint (author's final draft)

LoRa is a popular communication technology used in Internet of Things (IoT) applications. Typically, in the LoRaWAN architecture, an end node periodically sends a LoRaWAN message to a gateway connected to the Internet. Recent works, however, showed the possibility of LoRa mesh networks where LoRa nodes communicate with each other. In this paper we present a monitoring system as a tool for observing the traffic and the operation of such LoRa mesh networks. The client side is implemented at the LoRa nodes which periodically send to a server detailed information about the nodes' in-and outgoing LoRa packets. The server visualizes the information through a dashboard. Thus the monitoring tool allows network administrators to further analyze such LoRa mesh networks., This work was partially supported by the Spanish Government under contracts PID2019-106774RB-C21, PCI2019- 111851-2 (LeadingEdge CHIST-ERA), PCI2019-111850-2 (DiPET CHIST-ERA)., Peer Reviewed, Postprint (author's final draft)

SynFull is a widely employed tool that generates realistic traffic patterns for the performance evaluation of a NoC. In this work, we identify the main limitations of SynFull: high variability and long simulation time and also that these limitations increase when SynFull is integrated with RTL designs. SynFull-RTL employs a statistical approach, simulating each application macro-phase only once and averaging according to its probability of occurrence and the measured traffic load. SynFull-RTL obtains higher accuracy than the original version and reduced variability, with observed 40× reduction in simulation time and resources. A use-case with ProSMART validates the results., This work has been supported by the Spanish Science and Technology Commission under contract PID2019-105660RB-C22 and the European HiPEAC Network of Excellence. Enrique Vallejo has been partially supported by the Ministry of Universities, Subprograma Estatal de Movilidad, grant number PRX21/00757. This work also received funding from the European Union Horizon 2020 research and innovation programme under grant agreements number 826647 (EPI) and 946002 (MEEP)., Peer Reviewed, Postprint (author's final draft)

Network modeling is a fundamental tool in network research, design, and operation. Arguably the most popular method for modeling is Queuing Theory (QT). Its main limitation is that it imposes strong assumptions on the packet arrival process, which typically do not hold in real networks. In the field of Deep Learning, Graph Neural Networks (GNN) have emerged as a new technique to build data-driven models that can learn complex and non-linear behavior. In this paper, we present RouteNet-Erlang, a pioneering GNN architecture designed to model computer networks. RouteNet-Erlang supports complex traffic models, multi-queue scheduling policies, routing policies and can provide accurate estimates in networks not seen in the training phase. We benchmark RouteNet-Erlang against a state-of-the-art QT model, and our results show that it outperforms QT in all the network scenarios., This publication is part of the Spanish I+D+i project TRAINER-A (ref.PID2020-118011GB-C21), funded by MCIN/ AEI/10.13039/501100011033. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA) and the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund., Peer Reviewed, Postprint (author's final draft)

Wide Area Networks (WAN) are a key infrastructure in today’s society. During the last years, WANs have seen a considerable increase in network’s traffic and network applications, imposing new requirements on existing network technologies (e.g., low latency and high throughput). Consequently, Internet Service Providers (ISP) are under pressure to ensure the customer’s Quality of Service and fulfill Service Level Agreements. Network operators leverage Traffic Engineering (TE) techniques to efficiently manage the network’s resources. However, WAN’s traffic can drastically change during time and the connectivity can be affected due to external factors (e.g., link failures). Therefore, TE solutions must be able to adapt to dynamic scenarios in real-time. In this paper we propose Enero, an efficient real-time TE solution based on a two-stage optimization process. In the first one, Enero leverages Deep Reinforcement Learning (DRL) to optimize the routing configuration by generating a long-term TE strategy. To enable efficient operation over dynamic network scenarios (e.g., when link failures occur), we integrated a Graph Neural Network into the DRL agent. In the second stage, Enero uses a Local Search algorithm to improve DRL’s solution without adding computational overhead to the optimization process. The experimental results indicate that Enero is able to operate in real-world dynamic network topologies in 4.5 s on average for topologies up to 100 links., This publication is part of the Spanish I+D+i project TRAINER-A (ref.PID2020-118011GB-C21), funded by MCIN/ AEI/ 10.13039/501100011033. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA) and the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund., Peer Reviewed, Postprint (published version)

Next-generation satellite systems require more flexibility in resource management such that available radio resources can be dynamically allocated to meet time-varying and non-uniform traffic demands. Considering potential benefits of beam hopping (BH) and non-orthogonal multiple access (NOMA), we exploit the time-domain flexibility in multi-beam satellite systems by optimizing BH design, and enhance the power-domain flexibility via NOMA. In this paper, we investigate the synergy and mutual influence of beam hopping and NOMA. We jointly optimize power allocation, beam scheduling, and terminal-timeslot assignment to minimize the gap between requested traffic demand and offered capacity. In the solution development, we formally prove the NP-hardness of the optimization problem. Next, we develop a bounding scheme to tightly gauge the global optimum and propose a suboptimal algorithm to enable efficient resource assignment. Numerical results demonstrate the benefits of combining NOMA and BH, and validate the superiority of the proposed BH-NOMA schemes over benchmarks., This work was supported in part by the Luxembourg National Research Fund (FNR) CORE projects ROSETTA (under grant C17/IS/11632107) and FlexSAT (under grant C19/IS/13696663)., Peer Reviewed, Postprint (published version)