Abstract Photonic neural networks have emerged as a promising solution to overcome limitations of traditional hardware for neuromorphic computations, offering advantages in bandwidth, latency, and power efficiency. However, their performance is constrained by the limited precision of analog photonic computing, which is affected by inherent noise sources such as thermal and shot noise, and distortions. These effects degrade the photonic neural network performance, reducing the bit resolution achievable in photonic hardware typically to 2–4 bits. Traditional quantization strategies fail to account for these noise contributions, resulting in a substantial accuracy loss during inference. This paper introduces an adaptive quantization method called Adaptive-Quantization Photonic-Aware Neural Network (AQ-PANN) to address the challenges posed by different noise sources in analog photonic hardware. The proposed method uses a learnable step size quantization scheme to achieve high accuracy and stability under varying noise levels, introducing a scheme that unifies quantization step adaptation with noise injection exactly where photonic distortions arise. This design incurs only a minor training-time overhead, as it involves learning a small number of per-layer quantization step sizes and does not affect inference. Experimental evaluations on three commonly used test datasets (MNIST, SVHN, and CIFAR-10) with different bit resolutions show the robustness of AQ-PANN. On MNIST, an accuracy drop of only 2% was observed from low to high noise levels in a 4-bit configuration, while traditional DoReFa quantization suffered a 29% drop. For the SVHN dataset, AQ-PANN obtained a mean accuracy of 92% under high noise with 4-bit quantization, outperforming DoReFa by over 45%. On CIFAR-10, AQ-PANN maintained close to 60% accuracy under high noise in the 4-bit configuration, whereas DoReFa and PACT both collapsed below 40%. These results highlight the effectiveness of AQ-PANN in sustaining model performance across different noise intensities, enabling practical photonic neural network deployment.

Federated Learning (FL) enables collaborative model training without sharing raw data, making it attractive for privacy-preserving applications at the wireless edge. However, when executed over real 5G networks, FL performance degrades due to uplink congestion, heterogeneous client capabilities, and intermittent connectivity. Most existing approaches attempt to mitigate these issues indirectly by optimizing clients (through adaptive participation, local training, or selection strategies) or by optimizing models (via pruning, quantization, or compression), but they ignore potential network bottlenecks. This paper introduces FLAG, an FL architecture that embeds innetwork aggregation directly into 5G gNodeBs, transforming the network into an active participant in the learning process. In particular, FLAG performs parameter aggregation at line rate within the 5G Service Data Adaptation Protocol layer and incorporates three mechanisms: Partial-Contribution Correction for loss-tolerant averaging, a timer-driven pipeline for real-time scheduling, and a deadline-based grouping strategy to mitigate stragglers. Experiments with realistic wireless emulation show that FLAG achieves up to 5.1× faster time-to-accuracy and maintains accuracy within 0.8% of a loss-free baseline, while reducing gNB-to-server bandwidth by aggregating pergNB rather than per-client. FLAG requires no modifications to clients or the parameter server, demonstrating how 5G-aware system design can make federated learning scalable, efficient, and resilient under real-world wireless conditions.

The demand for continuous patients monitoring in healthcare has increased significantly, originated mainly from the need for proactive and personalized treatments. In fact, traditional periodical monitoring often fails to detect rapid physiological changes, resulting in a risk especially for patients with chronic conditions or post-surgical needs. This paper presents a voice-driven mobile health monitoring application that enables daily tracking of six health parameters, addressing the limitations of traditional periodic monitoring. Designed to cater to diverse environments, including rural and urban areas, the app ensures reliable functionality storing data locally and synchronizing with a centralized database once connectivity is available. The app integrates seamlessly with Bluetooth devices and $4 \mathrm{G} / 5 \mathrm{G}$ networks, offering scalability. It also relies on the QUIC protocol for a more secure data transmission and provides a voice recognition feature for an easy collection of health parameters thus further enhancing user experience and contributing to the growing field of mHealth.

Diver safety remains a critical concern in underwater exploration, with monitoring physiological parameters being crucial for assessing diver health. This preliminary study explores the feasibility of utilizing deep learning techniques for skin segmentation and the estimation of vital parameters to enhance diver safety. We employed an Attention U-Net based network to segment regions of interest from video sequences collected at $\mathbf{Y}$ 40 The Deep Joy in Montegrotto Terme, Italy, before and after a 30 meter-depth dive. The deep network was trained on images from the Face and Skin Detection database for the segmentation task. We achieved an accuracy of $97 \%$ and an intersection over union of $89 \%$ on the test data. Additionally, we extracted imaging photoplethysmography (iPPG) signals from the selected skin area and estimated vital parameters. The proposed pipeline was tested on data acquired in an underwater test environment, with reference data collected via pulse oximeter. Preliminary results demonstrate the potential of the proposed network to improve diver safety by providing real-time insights into diver health.

Cellular Vehicle-to-Everything (C-V2X) is currently at the forefront of the digital transformation of our society. By enabling vehicles to communicate with each other and with the traffic environment using cellular networks, we redefine transportation, improving road safety and transportation services, increasing the efficiency of vehicular traffic flows, and reducing environmental impact. To effectively facilitate the provisioning of Cellular Vehicular-to-Network (C-V2N) services, we tackle the interdependent problems of service task placement and scaling of edge resources. Specifically, we formulate the joint problem and prove that it is not computationally tractable. To address its complexity we propose dhpg, a new Deep Reinforcement Learning (DRL) approach that operates in hybrid action spaces, enabling holistic decision-making and enhancing overall performance. We evaluated the performance of DHPG using simulations with a real-world C-V2N traffic dataset, comparing it to several state-of-the-art (SoA) solutions. DHPG outperforms these solutions, guaranteeing the 99th percentile of C-V2N service delay target, while simultaneously optimizing the utilization of computing resources. Finally, time complexity analysis is conducted to verify that the proposed approach can support real-time C-V2N services.

In the context of NextG Wireless Networks, addressing the challenges of wireless communication link reliability is paramount to ensure efficient Distributed Learning systems. However, many recent solutions have overlooked key challenges, such as packet-level losses and the impact of TCP retransmissions, which are crucial for the robustness of these systems. In this paper, we propose the integration of fountain codes into the distributed learning process to offer a robust mechanism to counteract packet loss. Specifically, we propose a cumulative strategy logic based on fountain codes specifically tailored for packet exchanges in Distributed Learning applications. Our evaluation shows that fountain codes significantly enhance the efficiency and reliability of distributed learning model updates under severe packet loss conditions, e.g., a packet reduction of ≈ 84% (≈ 60%) at the UE (gNB) side compared to traditional TCP methods when packet loss probability reaches 0.9 in Federated Learning context. However, under low packet loss scenarios, fountain codes computational overhead becomes non-negligible. These results highlight the potential of fountain codes to serve as a robust alternative to conventional communication protocols in distributed learning systems, particularly in environments characterized by unstable network conditions.

The adoption of disaggregated, virtualized, and open gNodeB in the next generation Radio Access Network offers benefits such as cost reduction and improved network performance. However, meeting specific 5G and beyond requirements, e.g., Ultra Reliable Low Latency Communications, requires offloading selected gNodeB functions onto accelerated hardware. This study proposes to implement a 5G Distributed Unit (DU)Radio Unit (RU) in a System-on-a-Programmable-Chip (SoPC) where the FFT of Orthogonal Frequency Division Multiplexing in uplink transmission is offloaded onto an FPGA. The proposed solution is programmable and pluggable, allowing flexibility in function implementation and integration into various devices. The experimental evaluation shows that the proposed SmartNIC-based DU-RU achieves $15 \times$ speedup in processing time when compared to a server’s CPU with FPGA-accelerated Low-PHY processing. In addition, it shows that employing highperformance off-chip memory leads to about $14 \%$ reduction in processing time compared to the use of an FPGA-internal block memory.

Next-Generation Radio Access Networks (NG-RANs) aim to facilitate high data rates, low-latency applications, and dense mobile connectivity — benefit from the integration of Artificial Intelligence and Machine Learning (AI/ML) to enhance performance and efficiency. Nevertheless, the dynamic service demands within NG-RAN (namely Open RAN) lead to AI/ML performance degradation known as drift, resulting in violations of Service Level Agreements (SLA) and issues like over-or under-provisioning of resources. Detecting and adapting to drift becomes crucial to meet the diverse requirements of intelligent networks. Due to frequent retraining, the existing threshold and classifier-based approaches have potential disadvantages such as SLA violations and resource inefficiency. This paper introduces a novel approach that exploits the Generative Adversarial Network (GAN) architecture to determine the drift and anomaly. The proposed approach is evaluated for a throughput prediction use case over a real-time dataset and compared to the threshold and classifier-based approaches. The results show that the proposed approach outperforms the threshold and classifier-based approaches.

The implementation of zero-touch network and service management solutions in software defined optical networks requires the elaboration of detailed optical components’ information. However, such data can be elaborated by third parties. Thus, confidentiality issues may arise because providers are not willing to unveil their detailed information. This study proposes schemes based on dataset scrambling and unsupervised machine learning algorithms for soft failures detection in optical networks. A key aspect of the proposed scheme is the preservation of data confidentiality, that refers, in this context, to safeguard the detailed information of optical components while still enabling effective failure detection. The performance of six different clustering algorithms have been experimentally evaluated in a laboratory testbed. The results reveal that certain algorithms, while working in a confidentiality preserving scheme, perform very well in clustering different states (i.e., working and faulty states) of the network.

The adoption of Time Division Duplexing (TDD) in 5G not only facilitates efficient spectrum utilization but also paves the way for the implementation of innovative schemes that leverage the unique characteristics of the TDD patterns. For example, Cooperative Dynamic Bandwidth Allocation (CO DBA) exploits information about the TDD pattern to reduce latency in Time Division Multiplexing Passive Optical Networks (TDM-PON)-based fronthaul. In this paper, we harness the information about 5G TDD patterns alongside the programmability offered by Open Radio Access Network (O-RAN) and Software Defined Optical Access Networks (SDOANs) for a different purpose: enhancing x-haul energy efficiency. The scheme we propose seeks to minimize the energy consumption of PON-based x-haul networks by dynamically adjusting the operational states of selected subsystems within the Optical Network Terminal (ONT) in coordination with the TDD patterns. Preliminary simulation results show that up to 80% energy savings can be achieved with the proposed cooperative technique.

Imaging photoplethysmography (iPPG) has emerged as an alternative, contactless solution for the monitoring of physiological parameters. Applying iPPG in real-life scenarios, particularly in dynamic illumination with subject movement, presents a significantly challenging task. This paper introduces an iPPG-based non-invasive monitoring method for estimating heart rate while maintaining the privacy of the patient. The deep architecture consists of time-distributed convolutional neural networks and long short-term memory layers, trained on the PURE dataset. All images undergo a series of pre-processing steps, and the segmented region of interest is then provided as input to the deep architecture. The proposed methodology is evaluated using Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), achieving an RMSE of 3.52 and an MAE of 3.79 on test data derived from the PURE dataset, demonstrating comparable performance with other state-of-the-art algorithms. Furthermore, we evaluated the performance on video sequences acquired under dynamic conditions.

NextG wireless will heavily rely on Federated Learning (FL) applications to learn context-aware AI solutions from the massive amount of generated data. Ensuring the reliability of wireless links for such applications is paramount, especially for FL where packet loss can severely hamper performance and efficiency. Traditional approaches fall short under the high packet loss characteristics of wireless networks. This demo shows how the integration of Fountain Codes (FC) into the FL process can bring notable improvements in packet transmission efficiency, especially under high packet loss conditions.

This demonstration shows the effectiveness of a hierarchical Software-Defined approach in coordinating Radio Access Network (RAN) and Passive Optical Network (PON)-based RAN transport to proactively scale virtual Distributed Units (vDUs) / Radio Units (RUs) and to jointly reconfigure the mid-haul transport.

This paper investigates the performance of scalable user-centric (UC) distributed massive multiple-input multiple-output (D-mMIMO) systems with multiple central processing units (CPUs), commonly called cell-free mMIMO. Specifically, a framework incorporating processing capacity and inter-CPU communication constraints is proposed. Two methods are presented for limiting the number of radio units (RUs) serving each user equipment (UE). The first method is performed by the CPUs, while the second one is implemented at the UEs and RUs. Both methods prevent the computational complexity (CC) for channel estimation and precoding signals from increasing with the number of RUs. The backhaul signaling demands are presented and modeled, and it is considered that each CPU can serve only a restricted number of UEs managed by other CPUs to mitigate inter-CPU communication. Two strategies to adjust the RU clusters according to the network implementations are also proposed. We compare the proposed approaches with a traditional scalable UC system. Simulation results reveal that the proposed techniques allow UC systems to keep their spectral efficiency (SE) under minor degradation while reducing the CC by 98% and improving energy efficiency (EE). Besides, managing inter-CPU communication controls backhaul traffic effectively, and RU cluster adjustments further reduce CC.

This paper investigates the performance of scalable user-centric (UC) distributed massive multiple-input multiple-output (D-mMIMO) systems, widely known in the literature as cell-free mMIMO, with limited processing capacity. Specifically, it is assumed that the computational complexity (CC) of performing channel estimation and precoding signals does not increase with the number of access points (APs). In this regard, it is considered that each user equipment (UE) can only be associated with a finite number of APs. Moreover, a method is proposed for adjusting the AP clusters according to the network implementation, i.e., centralized or distributed. We compare the proposed approaches with a scalable UC system that does not perform AP cluster adjustment and does not prevent the processing demands from growing with the number of APs. Simulation results reveal that UC systems can keep the spectral efficiency (SE) under minor degradation even if the processing capacity is limited, reducing the CC by up to 96%. Besides, the proposed method for adjusting the AP cluster leads to further reductions in CC.

The widespread connectivity provided by fixed and mobile communication technologies can facilitate the utilization of e-health and mobile-health services that are radically changing the way healthcare may be provided. Such services are particularly important for inner areas, where it is difficult to guarantee physical availability and proximity of services while ensuring the sustainability of the system. However, e-health and m-health service deployment requires careful planning, as connectivity in inner areas can be spotty. This paper reports how multiple e/m-health solutions have been planned in some inner areas of Tuscany within the PROXIMITY-CARE project. In particular, it introduces a newly developed QGIS-based analysis tool, which allows correlating connectivity data with patient needs and healthcare facilities location. Moreover, the paper presents a tele-tutoring system for the emergency service and a mobile application to monitor patient vital parameters. Such tools facilitate the maximization of the reached population, thus improving benefits for the patients.

Beyond fifth-generation (B5G) networks (namely 6G) aim to support high data rates, low-latency applications, and massive machine communications. Integrating Artificial Intelligence (AI) and Machine Learning (ML) models are essential for addressing the network’s increasing complexity and dynamic nature. However, dynamic service demands of B5G cause the AI/ML models performance degradation, resulting in violations of Service Level Agreements (SLA), over-or under-provisioning of resources, etc. To address the performance degradation of the AI/ML models, retraining is essential. Existing threshold and periodic retraining approaches have potential disadvantages such as SLA violations and inefficient resource utilization for setting a threshold parameter in a dynamic environment. This paper presents a novel algorithm that predicts when to retrain AI/ML models using an unsupervised classifier. The proposed predictive approach is evaluated for a Quality of Service (QoS) prediction use case on the Open RAN Software Community (OSC) platform and compared to the threshold approach. The results show that the proposed predictive approach outperforms the threshold approach.

This paper investigates the performance of downlink (DL) pilot-based training to estimate the effective channel in user-centric cell-free massive multiple-input multiple-output (MIMO) networks. An algorithm for DL pilot assignment is proposed based on the level of interference between each user equipment (UE). It is proposed a refinement method for access point (AP) selection that controls the maximum AP cluster size of UEs. The strategy aims to control the maximum number of APs serving each UE to reduce the disparities among the AP cluster sizes. DL pilot-based training is compared with the blind, perfect and statistical channel state information (CSI) methods, assuming different precoding techniques, AP selection schemes, and the presence of pilot contamination. Our results demonstrate the following: (i) the proposed DL pilot assignment algorithm outperforms the baseline solutions; (ii) the proposed AP selection refinement method can improve the energy efficiency up to 86.6% without compromising the spectral efficiency; and (iii) DL pilot-based estimation reduces the normalized mean-square error significantly compared with blind and statistical CSI methods.

Ultra Reliable Low Latency Communications (URLLC) scenarios require very low latency and high reliability, imposing an optimization of every aspect of 5G data processing, transmission, and networking. Artificial Intelligence (AI)-based tools can be helpful resources in this context, enhancing multiple functionalities, from network resource allocation to network security. In this paper we propose a solution placed at the next generation eNB (gNB)-Central Unit (CU) level, relying on Neural Networks (NNs), capable of classifying incoming packets. The developed system increases the security of 5G and B5G architectures, protecting the 5G Core (5GC) from potential attacks. To comply with URLLC requirements on latency, we present an architecture leveraging photonic hardware to speed-up NN computations. The proposed solution, namely Photonic-Aware Neural Network (PANN), complies with physical layer constraints raised by photonic analog computing and can achieve high throughput and time-of-flight latency. The classification performance of the devised PANN model has been assessed through simulation on the distilled Kitsune dataset, suited for 5G scenarios. The experiments proved that PANN significantly lowers the chance of transmitting malicious packets to the 5GC with a classification performance increasing with the bit resolution supported by the analog photonic physical layer.

The 5G Radio Access Network (RAN) architecture provides a split option, whereby a gNodeB Central Unit (gNB-CU) is connected to one or more gNB-Distributed Units (gNB- DUs). The CU is in turn connected to the 5G Core Network (CN) and its functions can be virtualized through software containers. This demonstration showcases live migration of a containerized Central Unit (CU) component in a Cloud-native 5G network without loss of service. In terms of resiliency, virtual function live migration can circumvent the failure of the server hosting the gNB-virtualized CU (gNB-vCU) that would otherwise cause an interruption of user-plane (UP) traffic and disconnection of User Equipment (UE). The proposed gNB-vCU container live migration technique reduces the end-user service temporary downtime by 50% when compared to the traditional backup/restore option.

The advent of Open Radio Access Network (O-RAN) technology enables intelligent edge solutions for base stations in beyond 5G (B5G) networks. O-RAN Working Group 2 (WG2) focuses on the architecture and specifications of AI/ML workflows, allowing AI/ML applications in O-RAN environments to meet different QoS requirements for different use cases over varying time periods. This study shows the technical challenges in mapping AI/ML functionalities at Near-Real Time (RT) RAN Intelligence Controller (RIC) and/or Non-RT RIC for closed loop control-based resource adaptation in O-RAN. We also present a drift-based solution to avoid performance violations if there is decay in prediction accuracy. Results show that drift-based solution outperforms offline models.

Disaggregated, virtualized, and open next-generation eNodeB (gNB) could bring several benefits to the Next Generation Radio Access Network (NG-RAN) by enabling more market competition and customer choice, lower equipment costs, and improved network performance. This can be achieved through gNB-central unit (CU)-control plane (CP), gNB-CU-user plane (UP) and gNB-distributed unit (DU) separation, CU and DU function virtualization, and zero touch RAN management and control. However, to achieve the performance required by specific foreseen 5G usage scenarios (e.g., Ultra Reliable Low Latency Communications — URLLC), offloading selected disaggregated gNB functions into an accelerated hardware becomes a necessity. To this aim, this study proposes the implementation of 5G DU Low-PHY layer functions into an FPGA-based SmartNIC exploiting the Open Computing Language (OpenCL) framework to facilitate the integration of accelerated 5G functions within the mobile protocol stack. The proposed implementation is compared against (i) a CPU-based OpenAirInterface implementation, and (ii) a GPU-based implementation of IFFT exploiting clfft and cufft libraries. Experimental results show that the different optimization techniques implemented in the proposed solution reduce the Low-PHY processing time and the use of FPGA resources. Moreover, the GPU-based implementation of the cufft and the proposed FPGA-based implementation have a lower processing time and power consumption compared to a CPU-based implementation for up to two cores. Finally, the implementation in a SmartNIC reduces the delay added by the host-to-device communication through the Peripheral Component Interconnect Express (PCIe) interface, considering both functional split options 2 and 7-1.

Abstract Photonics-based neural networks promise to outperform electronic counterparts, accelerating neural network computations while reducing power consumption and footprint. However, these solutions suffer from physical layer constraints arising from the underlying analog photonic hardware, impacting the resolution of computations (in terms of effective number of bits), requiring the use of positive-valued inputs, and imposing limitations in the fan-in and in the size of convolutional kernels. To abstract these constraints, in this paper we introduce the concept of Photonic-Aware Neural Network (PANN) architectures, i.e., deep neural network models aware of the photonic hardware constraints. Then, we devise PANN training schemes resorting to quantization strategies aimed to obtain the required neural network parameters in the fixed-point domain, compliant with the limited resolution of the underlying hardware. We finally carry out extensive simulations exploiting PANNs in image classification tasks on well-known datasets (MNIST, Fashion-MNIST, and Cifar-10) with varying bitwidths (i.e., 2, 4, and 6 bits). We consider two kernel sizes and two pooling schemes for each PANN model, exploiting $$2\times 2$$ 2 × 2 and $$3\times 3$$ 3 × 3 convolutional kernels, and max and average pooling, the latter more amenable to an optical implementation. $$3\times 3$$ 3 × 3 kernels perform better than $$2\times 2$$ 2 × 2 counterparts, while max and average pooling provide comparable results, with the latter performing better on MNIST and Cifar-10. The accuracy degradation due to the photonic hardware constraints is quite limited, especially on MNIST and Fashion-MNIST, demonstrating the feasibility of PANN approaches on computer vision tasks.

The deployment of 5G and network slicing has challenged the current network management requirements, triggering the need for programmable and software-driven architectures. Thus, automated real-time fault management for self-managed networks with machine learning and artificial intelligence at the forefront has become necessary. This is especially the case of optical communication systems, accountable for most of the data traffic worldwide. This study introduces the application of a novel failure detection and localization framework capable of forecasting failures in optical systems based on an unsupervised learning strategy. In this approach, the Long- and Short-term Time-series Network (LSTNet) is exploited for modeling the normal behavior of optical systems. Then, failure conditions are properly forecast without explicitly training the model for such cases, easing the data acquisition process. Later, forecast values and actual measurements from optical equipment are used to derive an outlier detection method to detect and locate failures to improve the decision-making process at the network orchestrator. Laboratory experiments comparing the proposed approach with the Recurrent and Long Short-Term Memory models in terms of failure detection and forecasting performance show that using the LSTNet reduces the mean squared errors in 95% for unseen data, indicating robustness and suitability for real-world environments.

Network slicing plays a key role in the 5G ecosystem for vertical industries to introduce new services. However, one widely-recognized challenge of network slicing is to provide traffic isolation and concurrently satisfy diverse performance requirements, e.g., bandwidth and latency. In this work, we showcase the capability to retain these two goals at the same time, via extending the 5Growth baseline architecture and designing a new data-plane pipeline, i.e., virtual queue, over the P4 switch. To demonstrate the effectiveness of our approach, a proof-of-concept is presented serving different service requests over a mixed data path, including P4 switches and Open vSwitches (OvSs).

Empowered bynetwork softwarization, 5G systems have become the key enabler to foster the digital transformation of the vertical industries by expanding the scope of traditional mobile networks and enriching the network service offerings. To make this a reality, we propose anautomationsolution for vertical services provisioning and hierarchical Service Level Agreement (SLA) management.Service scalingis one of the most essential operations to adapt the service deployments and resource allocations to ensure SLA fulfilment. Three different scaling levels are addressed in this work: application-, service- and resource-level. We have implemented our solution in a proof-of-concept of a virtualized mobile network platform, spanning over three geographically-distributed sites. To evaluate our solution, we leverage field tests, focusing onautomotive vertical servicescomprising a mission-critical application (collision-avoidance) and an entertainment one (video streaming). The results demonstrate the excellent performance of our solution, and its ability to automatically deploy vertical services and ensure their SLAs through different levels of service scaling.

The Open Computing Language (OpenCL) is increasingly adopted for programming processors with reconfigurable hardware acceleration. The 5G telecommunication infrastructure, imposing strong latency constraints on the managed communications, may benefit from OpenCL-designed accelerated processing. This paper presents the first study to evaluate OpenCL hardware acceleration in the context of a 5G base station physical layer. The implementation and optimization process to accelerate the Orthogonal Frequency Division Multiplexing (OFDM) part of the 5G downlink is conducted on a high-end Field Programmable Gate Array (FPGA). We show that the proposed OpenCL implementation complies with the 5G processing timing requirements since the computation time is consistent with the present 5G deployment. However, to be suitable for 5G, the OpenCL platform must improve the data latency transfer between hardware and software. Moreover, a further enhancement for the OpenCL implementation is to improve the code by means of OpenCL optimization techniques. In this way, the performance can be further improved with respect to optimized software on vectorized high-end processors.

The fifth generation of mobile communications (5G) is expected to dramatically improve performance compared to preceding standards by offering very high bandwidths and low latencies. To provide this performance, heavy processing is required and must meet strong timing constraints. Reconfigurable computing, managing processing in software and exploiting reconfigurable hardware acceleration, is an innovative approach that should be considered for 5G for its capacity to combine high throughput and high flexibility. This paper presents a case study for Orthogonal Frequency Division Multiplexing (OFDM) computation reconfigurable offloading onto an Field Programmable Gate Array (FPGA). The implementation is based on Open Computing Language (OpenCL) that represents a versatile solution, as this language can be compiled for several architectures, provided that a Host+Accelerator structure is used. The objective of our study is to demonstrate that, by means of hardware offloading, the 5G architecture resources can reach high computational load, avoiding processing stalls and latency increase. Results show that around 15% of the software processing can be freed through hardware acceleration and reallocated to support other tasks.

The concept of network slicing in 5G infrastructures allows to deliver multiple virtual services over shared environments and fully customized based on vertical-driven requirements and target performance indicators. A key feature of 5G networks is the capability to dynamically re-optimize the allocation of computing and network resources, from the core up to the edge, to instantiate and manage different types of concurrent services over a shared infrastructure. In this demonstration, we present a network slicing and orchestration solution for vertical services in the media sector, where enhanced Mobile Broadband (eMBB) network slices are instantiated interconnecting physical and virtual functions, provisioned and configured on-demand using the concept of NFV Network Services. The eMBB slices are automatically dimensioned to match the requirements of the virtual Content Delivery Network (vCDN) service, e.g. in terms of number of target users, video quality and geographical coverage area. Arbitration and resource allocation schemas optimize the sharing of mobile communication services and virtual resources among concurrent media service instances, in compliance with the Service Level Agreement between network operators and vertical service providers.

Softwarization of mobile and optical networks facilitates the inter-working between control planes of the two domains, allowing a more efficient management of available resources. Radio resource utilization benefits from the centralization of mobile network functionalities with the application of high-order functional split options by fronthauling. However, higher-order options require larger bandwidth and lower latency in the fronthaul. Advanced mechanisms for the joint control of the access network represent the sole solution to support such fronthaul requirements. This paper proposes a new cooperation scheme to manage the adaptive flexible functional split in 5G networks conditioned to the resource availability in the optical access network. Results show that the proposed converged approach guarantees the optimal allocation of optical resources through a software defined wavelength and bandwidth allocation. The proposed scheme adapts to current traffic demand and simultaneously allows the mobile network to take advantage of the highest possible centralization of mobile network functions by leveraging flexible functional split adaptively compliant to the current optical traffic demand.

The utilization of Ethernet in the cellular terrestrial cloud radio access network (C-RAN) fronthaul is considered as a way for improving C-RAN network reconfigurability and efficiency in terms of both capital and operational expenditures. Moreover CPRI line bit rate dynamic reconfiguration may spare further capital expenditures by avoiding a peak traffic capacity allocation. A possible solution for introducing Ethernet in the fronthaul is the encapsulation of Common Packet Radio Interface (CPRI) over Ethernet. However, it must be assured that CPRI strict requirements on delay and jitter are still met. In this work the combined impact of encapsulating CPRI on Ethernet and of the dynamic CPRI line bit rate reconfiguration on delay and jitter is evaluated. The evaluation is both analytical and based on the pre-synthesis emulation of dynamic CPRI line bit rate reconfiguration. Results show that dynamic CPRI line bit rate reconfiguration can be achieved within about one millisecond. However, if a size-based encapsulation of CPRI over Ethernet is utilized, dynamic CPRI line bit rate reconfiguration might cause delay variations (i.e., jitter) up to few microseconds.

Fiber-wireless sensor networks (Fi-WSNs) composed of a hybrid fiber-wireless (FiWi) network enhanced with sensors will play a key role in supporting machine-to-machine (M2M) communications to enable a wide range of Internet of Things (IoT) applications, of which smart grids represent an important real-world example. This paper explores opportunities of designing an energy-efficient Fi-WSN based on EPON/10G-EPON, WLAN, wireless sensors, and passive fiber optic sensors as a shared communications infrastructure for broadband services and smart grids. A novel energy conservation scheme for sensor enhanced FiWi networks (ECO-SFiWi) is proposed to reduce the overall energy consumption. ECO-SFiWi maximizes energy efficiency by leveraging TDMA to schedule power-saving modes of EPON's optical network units, wireless stations, and wireless sensors and incorporate them into EPON's bandwidth allocation algorithm. To study the performance, a comprehensive energy saving model and a delay analysis of both FiWi traffic and sensor data based on M/G/1 queue modeling are presented. FPGA-based hardware emulation and demonstration are performed to verify the effectiveness of the proposed solution. Results provide deep insights into the tradeoff between energy savings and frame delays. Noticeably, ECO-SFiWi achieves significant amounts of energy saving, while maintaining low delay for FiWi traffic and sensor data under typical deployment scenarios.

Integrated fiber-wireless (FiWi) access networks aim at taking full advantage of the reliability and high capacity of the optical backhaul along with the flexibility, ubiquity, and cost savings of the wireless/cellular front-end to provide broadband services for both mobile and fixed users. In FiWi access networks, energy efficiency issues must be addressed in a comprehensive fashion that takes into account not only wireless front-end but also optical backhaul segments to extend the battery life of wireless devices and allow operators to reduce their OPEX, while not compromising quality of service (QoS). This paper proposes an energy conservation scheme for FiWi networks (ECO-FiWi) that jointly schedules power-saving modes of wireless stations and access points and optical network units to reduce their energy consumption. ECO-FiWi maximizes the overall network performance by leveraging TDMA to synchronize the power-saving modes and incorporate them into the dynamic bandwidth allocation (DBA) process. A comprehensive energy saving model and an M/G/1 queuing-based analysis of downstream and upstream end-to-end frame delays are presented accounting for both backhaul and front-end network segments. Analytical results show that ECO-FiWi achieves significant amounts of energy saving, while preserving upstream delay and incurring a low delay for downstream traffic.

In TWDM-PON, the utilization of tunable transceivers at the ONUs allows to dynamically reconfigure the OSU-ONU pairings to decrease, for example, the energy consumption at the OLT. However, if the utilized dynamic wavelength and bandwidth allocation (DWBA) does not take into account which ONUs are reallocated, the ONU tuning overhead can severely impact the average frame delay experienced by the tuning ONUs. The paper shows that, by alternating the ONUs that tune at each reconfiguration, the impact of ONU tuning overhead on average frame delay is fairly distributed among the ONUs.

When traffic carried by a Time and Wavelength Division Multiplexed (TWDM) PON is fluctuating, the network performance, both in terms of energy efficiency and delay, is heavily impacted by the decision of when network reconfiguration (i.e., switching OSUs from active mode to sleep mode and vice versa) is triggered. This paper provides hints on when to trigger a network reconfiguration is such scenario by considering also the ONU tuning time. Results show that reconfigurations shall follow as soon as possible traffic variations. However, if traffic variations are fast, the ONU tuning time impairs delay performance. Therefore, when traffic variations are fast, it is better not to reconfigure the network at the expense of a higher energy consumption.

We previously proposed and numerically analyzed a theoretical framework of an energy-efficient offline dynamic wavelength and bandwidth allocation (DWBA) algorithm designed for a delay-constrained time and wavelength division multiplexed passive optical network (TWDM-PON). This DWBA algorithm exploits the tunability and the sleep/doze capabilities of a 10 Gbps vertical-cavity surface-emitting optical network unit (10G-VCSEL-ONU) to improve the energy-savings at the OLT and the ONUs, respectively. In this work, using simulation results on the number of active wavelengths and the percentage of energy-savings, we verify the theoretical framework proposed in our previous study. Most importantly, we show that the average delay of upstream packets are not adversely affected by the proposed energy-saving mechanism and is kept below the specified maximum.

This paper proposes an advanced sleep-aware dynamic bandwidth allocation (ASDBA) scheme for 10G-EPONs that aims to maximize ONU energy efficiency with sleep mode. In the proposed ASDBA scheme, both upstream (US) and downstream (DS) transmissions are scheduled in the same transmission slot whose duration is minimized based on both DS and US bandwidth requests and the ONU transceiver is switched off outside the slot for saving energy. The ASDBA swaps the conventional order of control message exchange utilized in legacy 10G-EPON to convert the ONU idle time between a REPORT message and its replying GATE message into ONU sleep time. This enables the ONU to sleep continuously after sending a REPORT message until the beginning of the next transmission slot, further improving ONU energy-savings. Results show that the proposed scheme significantly saves ONU energy whilst incurring acceptable frame queuing delays.

This paper presents the experimental evaluation of an energy-efficient TDMA PON utilizing cyclic sleep technique. In particular, two methods for deciding when to trigger ONU sleep mode are evaluated: downstream (DS)-based triggering and Cooperative triggering. In the former method, the decision whether the ONU is switched to sleep is taken based on the DS traffic only. In the latter one, the decision is taken by considering both DS and upstream (US) traffic. The performance is assessed with constant sleep time as well as variable sleep time for different DS/US traffic ratios. Experimental results show that both triggering methods achieve significant energy savings. The Cooperative saves slightly less energy than the DS-based while providing better network performance, in terms of frame loss rate and frame delay, for both DS and US traffic.

The International Electrotechnical Commission (IEC) recently standardized several Industrial Ethernet solutions that introduce the fieldbus concepts within Ethernet based networks. In addition, the IEC 62439 standardized a set of redundancy management protocols, including the Media Redundancy Protocol (MRP). In this way, IEC standards provide a variety of Ethernet-based solutions for satisfying both temporal and redundancy management requirements of Industrial Area Networks (IANs). In this paper, after a detailed study and implementation of MRP, two factors are identified that have an important impact on the protocol performance: the offset time and the physical detection time. A method is then provided to calculate a threshold to the network recovery time. Finally, extensive simulations and experimental measurements are performed to accurately evaluate the effect of the aforementioned factors on the protocol performance.

In multi-domain multi-carrier networks the effective use of network resources shall be achieved while guaranteeing an adequate level of confidentiality and scalability. A candidate solution to perform effective multi- domain Traffic Engineering (TE) is based on a combination of (i) hierarchical routing and (ii) path computation procedures. Hierarchical routing identifies the domain sequence to cross while path computation computes the strict end to end path. In this multi-domain study we first propose a hierarchical instance of BGP (HBGP) dedicated to TE information only. Then we propose and evaluate the integration of HBGP with path computation procedures based on IETF PCE architecture. Simulation results show that the hierarchical HBGP-PCE architecture, compared to current routing solutions based on BGP only, significantly improves the overall network resource utilization. In addition, this study identifies the network scenarios in which the aforementioned HBGP-PCE features provide significant advantages. Finally, the experimental implementation of the proposed HBGP-PCE architecture in a network testbed composed of commercially available routers shows the viability of the solution in real networks.

In GMPLS-controlled wavelength switched optical networks (WSONs), the RSVP-TE signaling protocol is utilized to reserve resources during both lightpath provisioning and dynamic restoration. During restoration, a number of reservation instances are contemporarily triggered by the failure. In such dynamic conditions, resource contention is the main blocking source. Several distributed solutions to reduce the impact of resource contention have been proposed that utilize advanced signaling mechanisms extending the RSVP-TE protocol. Conversely, this paper proposes two centralized solutions based on the Path Computation Element (PCE) that is used to coordinate the dynamic restoration of disrupted lightpaths with the specific aim of reducing resource contentions. Simulation results show that the utilization of the PCE during restoration reduces the blocking at the expenses of an increased recovery time with respect to distributed solutions. In addition, if the PCE is utilized together with the aforementioned advanced signaling mechanisms, the blocking can be further reduced.

In the past few years, Ethernet Passive Optical Networks (EPONs) have rapidly gained large popularity in broadband access networks. However the current standard has no management protocols aiming at reducing power consumption. In this paper, we propose an Energy Management Mechanism (EMM) within the IEEE 802.3ah control scheme. The main idea is to switch Optical Network Units (ONUs) to sleep mode and determine a suitable wakeup time schedule at the Optical Line Terminal (OLT). In the proposed EMM there is a trade-off between maximizing the power saving and guaranteeing the network performance. We compare two types of downstream scheduling schemes, Upstream Centric Scheduling (UCS) and Downstream Centric Scheduling (DCS), which are different in the way they assign 'active' and 'sleep' states to ONUs. Simulation results in terms of energy consumption and queuing delay are shown for the EMM based EPON system.

The connection-oriented nature of IEEE 802.16 (WiMAX) protocol facilitates the handling of quality of service (QoS). In IEEE 802.16 (WiMAX) networks, QoS-guaranteed connections, also referred to as service flows, can be dynamically activated between the base station and the subscriber stations, by using a three-way handshake protocol, called dynamic service addition (DSA). However, the unreliability of the radio medium may require multiple retransmission of DSA messages, leading to a delayed or even unsuccessful activation of a service flow. This paper proposes the novel concept of Quality of Activation (QoA) to guarantee the performance of the DSA protocol. Unlike QoS concept, QoA concept aims at guaranteeing the quality of the DSA message transmissions, rather than that of data transmission. In this paper, QoA is defined in terms of signaling blocking and maximum latency and is achieved by determining a set of constraints that limit the range of DSA protocol parameters. The impact of QoA requirements on protocol parameters is evaluated for various channel quality scenarios and for different types of service flows, i.e., from delay-sensitive service flows to critical and delay-insensitive service flows.

If dynamic bandwidth-guaranteed connections between distributed services (e.g., grid services) are provisioned through a centralized system, the policy to serve connection requests might heavily impact both the success in and the time required for setting up user services (e.g., grid-enabled applications). In this paper, the implementation of a batch queue in the centralized system is proposed. By implementing different service policies for the queued requests, connections and, in consequence, user services can be set up with different guarantees. In this study, a bulk-service policy is proposed and implemented to maximize connection set up success. The experimental evaluation results show that the utilization of the proposed policy brings advantages in terms of percentage of accepted connection requests as the number of requests served in one batch increases. Moreover, the achieved improvement does not impact the time required to set up the connections because of the specific LSP set up procedures implemented in the utilized commercial routers.

Distributed application (e.g., grid-enabled application) performance is highly dependent on the information available when computational resources are chosen. A resource selection based on computational resource information complemented with network performance information has the potential to be optimal from the application performance viewpoint. This is particularly true for network-intensive distributed applications. This study proposes a recursive distributed topology discovery service (RD-TDS) that allows grid clients to retrieve network performance information (i.e., IP-level topology and link capacity) without the need of specific administrative privileges. The RD-TDS exploits a selected set of distributed beacons (i.e., measurement points) that recursively probe newly discovered nodes until no undiscovered nodes are found during an exploration step. The RD-TDS simulative and experimental evaluation confirms its expected qualities: a rapid and complete discovery of the network performance information with the utilization of a limited number of active beacons. In addition, the proposed method rationale can be easily applied to many current network exploration tools.

Wireless and wired networks have commonly evolved separately. However with the integration of quad-play services (i.e., mobile/fixed voice, video, and data) into a single network infrastructure the provisioning of end-to-end connections spanning both the wireless and the wired domains becomes an important issue. In this paper two solutions are proposed for implementing an integrated end-to-end signaling for establishing QoS-guaranteed connections in integrated WiMAX and MPLS based networks. The two solutions are based on the integration of WiMAX management messages and RSVP-TE signaling. Their main difference consists in the way flow WiMAX message and RSVPTE signaling are triggered and coordinated.

Global Grid Computing goal is to connect heterogeneous computational resources belonging to the same Virtual Organization (VO) through the Internet to form a single, more powerful virtual computer. However, to fulfill this goal it is necessary to develop services that provide the virtual computer with the same functionalities of individual end systems, such as security, interprocess communication, and resource management.

The IEEE 802.16 standard for wireless metropolitan area networks allows communications between subscriber stations (SS) through the utilization of the optional Mesh operation mode. In mesh mode, the network base station can centrally schedule transmissions by routing packets on the links of a tree rooted at the BS and covering the mesh network. In this study, strategies for improving the reliability of the message broadcasting from the BS to the SSs are proposed and evaluated. By using the, so called, mesh assisted broadcasting (MAB) strategies, SS may exploit the higher connectivity degree offered by the mesh network during reception and may optionally cooperate with other SSs during the retransmission. MAB strategy performance is evaluated analytically and through simulations. Due to the complexity of the computational problem, analytical models are developed only for the MAB strategies for which the probability that an SS correctly receives a broadcast message can be obtained in closed form. Analytical and simulation results show that MAB strategies may turn the inherent broadcast nature of the radio medium into an increase of broadcasting reliability with respect to the tree-based broadcasting strategy defined in the IEEE 802.16 standard.

In grid computing the need for collecting information about both distributed computational resources and network topology and performance is constantly growing. Several tools are currently under development to provide such information. They are based either on a centralized or a distributed architecture. Distributed tools are commonly based on IP-level application- oriented network metrology. The measurements are done by means of beacons running in some network nodes (e.g., grid hosts) and collecting the required information. However, the number of utilized beacons might heavily impact the final result, i.e. the discovered topology and the collected performance information. In this study a model is developed to estimate the percentage of discovered links provided that a specific number of beacons is placed in the network. The model is developed for Erdos-Renyi (ER) graph network models but it can be applied also to other networks. Numerical evaluation shows that the model closely approximate the percentage of discovered links obtained through simulation for ER networks. For other theoretical and real networks the model overestimates the percentage of discovered links. However experimental results show that for networks with realistic average nodal degree the overestimate is less than 20%.

This paper presents the architecture, some specific supporting functions and an experimental validation of a new functional plane, namely the service plane, for realizing an added-value service provisioning (e.g., grid connectivity) for telecommunication operators. First, it is shown as the service plane can be a viable solution for decoupling service and transport development, by masking the transport-related implementation details from the abstract request of a service by a customer or by a qualified application. To this purpose, the service plane exports a high-level interface for supporting application-initiated invocation of QoS-enabled virtual private networks (VPN) or connection-less services. As a significant use case for a grid user, a VPN set-up through the service plane is experimentally demonstrated. Second, some of the main functions that the service plane should support are presented in detail and experimentally assessed, namely a centralized topology discovery service (C-TDS) and path computation service (PCS). As an example, from a grid user perspective, the C-TDS can provide up-to-date information on the grid topology according to various levels of abstraction (physical topology, MPLS topology, and logical topology). Several techniques for the grid topology discovery and various update policies are investigated. PCS elaborates upon the logical topology obtained by TDS and runs linear programming (LP) formulations to identify optimal traffic engineering solutions according to specific objective functions. The combination of C-TDS and PCS represents an an enhanced level of network- awareness in the (network) middleware supporting global grid computing (i.e., grid computing in wide area networks). Experiments performed on IP/MPLS metropolitan network based on commercial routers exhibit a topology delivery performance within a time span in the order of a few seconds and a PCS operation in the order of ten seconds.

Next generation GMPLS networks will be characterized by domains of transparency, in which the end-to-end optical signal quality has to be guaranteed. Currently GMPLS does not take into account the evaluation of physical impairments. Thus just limited size domains of transparency, where physical impairments can be neglected, are practically achievable. This study utilizes GMPLS signaling protocol extensions to encompass the optical layer physical impairments. The proposed approach allows to detect during the signaling phase whether lightpaths cannot be set up because of unacceptable optical signal quality. In this case successive set up attempts are performed, selecting the alternative routes with three schemes which exploit the feedback information of the signaling messages. Numerical results show that the proposed extensions are able to significantly decrease the lightpath blocking probability due to physical impairments in both static and dynamic conditions.

In modern communication networks different traffic classes require different quality of service (QoS) guarantees. It is therefore important for the benefit of both communication network users and network providers to meet each traffic class requirement while minimizing the utilized network resources. In this paper the failure-aware idle protection capacity reuse concept, FAIR for short, is presented and evaluated in a network scenario with two connection classes: high and low class lightpaths. High class lightpaths require full reliability against any single failure and are therefore shared path protected. Low class lightpaths might tolerate not be recovered after failure and are therefore unprotected. If low class lightpaths are disrupted by a link failure, they resort to best effort dynamic restoration. The FAIR concept is based on the utilization of idle (i.e., unutilized in the specific failure scenario) high class connection protection resources to recover disrupted low class connections. Numerical results show that the FAIR concept offers a simple yet efficient way to maximize network resource utilization while preserving connection reliability requirements without increasing control protocol overhead. The dynamic restoration of disrupted low class connections on free and idle resources brings a twofold advantage. On one hand it allows to significantly improve low class connection likelihood to be recovered. On the other hand, given a specific low class connection restoration blocking probability threshold, it increases the number of accepted low class connections during the provisioning phase. In addition the dependence of the aforementioned advantages on the network average nodal degree is negligible

Reliability of data exchange is becoming increasingly important. In addition, applications may require multiple degrees of reliability. The concept of differentiated reliability (DiR) was recently introduced in [A. Fumagalli and M. Tacca, January 2001] to provide multiple degrees of reliability in protection schemes that provision spare resources. With this paper, the authors extend the DiR concept to restoration schemes in which network resources for a disrupted connection along secondary paths are sought upon failure occurrence, i.e., they are not provisioned before the fault. The DiR concept is applied in two dimensions: restoration blocking probability i.e., the probability that the disrupted connection is not recovered due to lack of network resources - and restoration time - i.e., the time necessary to complete the connection recovery procedure. Differentiation in the two dimensions is accomplished by proposing three preemption policies that allow high priority connections to preempt resources allocated to low priority connections. The three policies trade complexity, i.e., number of preempted connections, for better reliability differentiation. Obtained results indicate that by using the proposed preemption policies, it is possible to guarantee a significant differentiation of both restoration blocking probability and restoration time. By carefully choosing the preemption policy, the desired reliability degree can be obtained, while minimizing the number of preempted connections.