With the rapid global adoption of fifth-generation (5G) mobile telecommunications, the demand for highly flexible private networks has surged. A key beyond-5G feature is network slicing, where the 3rd Generation Partnership Project (3GPP) defines three main use cases: massive Machine-Type Communications (mMTC), enhanced Mobile Broadband (eMBB), and Ultra-Reliable Low-Latency Communications (URLLC), along with their associated management functions. Similarly, the European Telecommunications Standards Institute (ETSI) provides the Zero-Touch Network and Service Management (ZSM) standard, enabling operation without human intervention. However, current technical documents lack definitions for end-to-end (E2E) management and integration across domains and subnet instances. We present a Network Slice as a Service Platform (NASP) that is agnostic to 3GPP and non-3GPP networks, addressing this gap. The NASP architecture comprises (i) onboarding requests for new slices at the business level, translating them into definitions of physical instances and interfaces among domains, (ii) a hierarchical orchestrator coordinating management functions, and (iii) communication interfaces with network controllers. Our NASP prototype is developed based on technical documents from 3GPP and ETSI, analyzing design overlaps and gaps across different perspectives. Results demonstrate the platform’s adaptability in handling diverse requests via the Communication Service Management Function. Evaluation indicates that the Core configuration accounts for 68% of the time required to create a Network Slice Instance. Tests reveal a 93% reduction in data session establishment time when comparing URLLC and Shared scenarios, i.e., an eMBB-type slice that reuses existing control-plane Network Functions to represent a best-effort provisioning baseline. Finally, we present cost variations for operating the platform with the orchestration of five and ten slices, showing a 112% variation between Edge and Central deployments.

The future mobile network schedulers have the complex mission of distributing radio resources among various applications with different requirements. The radio access network (RAN) slicing enables the creation of different logical networks by using dedicated resources for each group of applications. In this scenario, the radio resource scheduling (RRS) is responsible for distributing the radio resources among the slices to fulfill their requirements. Several recent studies have proposed advances in machine learning-based RRS. However, these works often evaluate their models under limited scenarios and with minimal slice diversity, raising concerns about their real-world applicability. The generalization capabilities of these models remain uncertain without rigorous testing across diverse network conditions and slice configurations, which may hinder their effectiveness upon deployment in operational networks. This paper proposes an intent-based RRS using multi-agent reinforcement learning in a RAN slicing context. The proposed method protects high-priority slices when the available radio resources are insufficient, using transfer learning to reduce the number of required training steps. The proposed method and baselines are evaluated in different network scenarios that comprehend combinations of different slice types, channel trajectories, number of active slices and users' equipment (UEs), and UE characteristics. The proposed method outperformed the baselines in protecting slices with higher priority, obtaining an improvement of 40% and, when considering all the slices, obtaining an improvement of 20% in relation to the baselines.

In this article, we propose a novel formulation that jointly considers the Virtualized Network Function (VNF) placement at the Radio Access Network (RAN) nodes and the resource allocation for the Transport Network (TN) of sliced and disaggregated RANs. Unlike most works in the literature that address these network aspects separately, we propose a joint approach. Our proposal ensures an end-to-end delay bound for the Ultra-Reliable and Low-Latency Communications (URLLC) use case in an Industry 4.0 scenario, while simultaneously considering the number of admitted User Equipments (UEs), the transmission rate allocation per slice, the functional split of RAN nodes, and the routing paths in the TN. We use Network Calculus (NC) theory to calculate delay along the TN connecting disaggregated RANs deploying network functions at the Radio Unit (RU), Distributed Unit (DU), and Central Unit (CU) nodes. The maximum end-to-end delay is modeled as a constraint in the optimization-based formulation that maximizes the number of admissible flows meeting strict Quality of Service (QoS) requirements, while also taking into account the deployment and operational costs of disaggregated RANs. In this approach, we propose a strategy based on real data from one of the world’s leading Mobile Network Operators (MNOs) to derive coherent estimations to the weights of the proposed objective function. The optimization model leverages a Flexible Functional Split (FFS) approach to provide a new degree of freedom to the resource allocation strategy. Simulation results reveal that, due to its non-linear nature, there is no trivial solution to the proposed optimization problem. Simulation results also show that our proposal guarantees a maximum delay for URLLC use cases in Industry 4.0 while satisfying bandwidth requirements for enhanced Mobile Broadband (eMBB) services.

This demonstration addresses a key open challenge in Open Radio Access Network (O-RAN) deployments: how to intelligently allocate Transport Network (TN) resources to ensure low-latency for mission-critical applications. The demo emulates a Smart Factory scenario where the time-sensitive control traffic of robotic arms competes with industrial camera broadband video streams. We propose an intelligent transport controller that combines network slicing, Adaptive Neuro-Fuzzy Inference System (ANFIS), and Federated Learning (FL) to dynamically prioritize traffic per slice. The architecture uses $\mathbf{P 4}$ switches for local queue monitoring and real-time resource scheduling. The integration with the O-RAN disaggregated stack is based on Open Air Interface (OAI). Experimental results demonstrate valuable load balancing and buffer occupation reduction in the O-RAN midhaul.

The advent of fifth-generation mobile networks (5G) has brought transformative improvements in data transfer speeds, latency reduction, and device connectivity, enabled by the softwarization, virtualization, and disaggregation of network functions. Open-source projects have been instrumental in facilitating software-driven 5G deployments on standard hardware, accelerating the development of private 5G networks, test environments, and advancing research and innovation in the domain. Despite these advancements, the current maturity of open-source solutions remains inadequate for high-performance, large-scale 5G deployments. Existing 5G performance testing approaches primarily adopt a black-box perspective, which fails to capture the intricate interactions among core 5G components, thereby hindering the identification of performance bottlenecks. Additionally, testing is often limited to small-scale environments, and publicly available datasets from real-world deployments are scarce. To overcome these challenges, this paper introduces an observability-driven methodology for testing$5G$core network performance, offering deeper insights into the internal interactions of network components. We contribute to research and innovation in 5G and beyond field by (i) enabling the simulation of varying user arrival rates, (ii) releasing a public dataset capturing data of 5G core components interactions, logs, and performance metrics, and (iii) providing a comparative analysis of two popular open-source 5G core implementations based on our observability-oriented methodology.

Extended reality (XR) services form the basis of various innovative applications. These new applications are expected to run on mobile devices with limited computational and energy capabilities. At the same time, XR services should fulfill some expectations regarding data rates and end-to-end latency to guarantee an uninterrupted user experience. In this context, offloading intensive computation to the edge is particularly advantageous for mobile devices, enabling access to more capable processing hardware. Current studies on computational offloading focus mainly on the compute and network continuum formed between edge and cloud and realized mostly through containers. Conversely, this demonstration showcases a computational offloading framework that allows the expansion of XR services functionality from the connected mobile device to an edge computing environment. The presented framework is based on a portable, lightweight, and secure WebAssembly runtime and uses open technology implementations. We demonstrate that the developed framework allows significant performance improvements in a Yolo object detection application and reduces heat generation on mobile devices.

The fifth generation of mobile networks (5G) has introduced significant advancements in data transfer rates, latency, and device density, empowered by softwarization, virtualization, and disaggregation of network functions. Open-source initiatives have played a crucial role in enabling software-based 5G network deployments on commodity hardware, driving the growth of private 5G networks and testbeds, and fostering research and innovation in the field. However, the maturity level of such open-source technologies is still insufficient for highly performant large-scale deployments. Current 5G performance testing methodologies rely predominantly on a black-box approach, which overlooks the complexity of interactions between 5G core components, limiting the ability to explain performance bottlenecks. To address this limitation, this paper proposes a new observability-driven methodology for 5G core network performance testing, which improves visibility into internal component interactions. We contribute to research and innovation in 5G and beyond field by (i) implementing testing workloads that are able to simulate varying user arrival rates and (ii) integrate a set of cloud-native observability tools to allow white-box performance evaluation of the 5G core.

The Radio Access Network (RAN) is the segment of cellular networks that provides wireless connectivity to end-users. The O-RAN Alliance has been transforming the RAN industry by proposing open RAN specifications and the programmable Non-Real-Time and Near-Real-Time RAN Intelligent Controllers (Non-RT RIC and Near-RT RIC). Both RICs provide platforms for running applications called rApps and xApps, respectively, to optimize the RAN behavior. We investigate the disaggregation of the Near-RT RIC into components that meet stringent latency requirements while presenting a cost-effective solution. For example, the O-RAN Signalling Storm Protection requires the Near-RT RIC to support end-to-end control loop latencies as low as 10 ms. We propose the novel RIC Orchestrator (RIC-O) that optimizes the deployment of the Near-RT RIC components across the cloud-edge continuum. Edge computing nodes often present limited resources and are expensive compared to cloud computing. Performance-critical components of Near-RT RIC and certain xApps should run at the edge while other components can run on the cloud. Furthermore, RIC-O employs an efficient strategy to react to sudden changes and re-deploy components dynamically. The proposal is evaluated both analytically and through real-world experiments in an extended Kubernetes deployment implementing RIC-O and the disaggregated Near-RT RIC.

Beyond 5G (B5G) and 6G networks must offer network slicing as a service to support disruptive applications using mobile network infrastructures. Moreover, network slicing as a service should enable the orchestration and management of disaggregated radio access networks (RAN), i.e., it allows the automation and abstraction of network configurations composed of physical and virtualized components such as defined by the 3rd Generation Partnership Project (3GPP) and Open RAN (O-RAN) Alliance. Network slicing must reach the part of the network between the radio units and the distribution units, i.e., the fronthaul network. Fronthaul is essential and diverse in 5G and B5G networks and can be composed of point-to-point or multipoint connections. In this context, the literature presents several works investigating the problems of network slicing and disaggregated networks. However, no survey explores current works integrating network slicing in disaggregated networks, specifically in the fronthaul network. This article surveys network slicing on disaggregated networks and its potential use in point-to-point and multipoint fronthaul infrastructures. We cover the state-of-the-art by analyzing four fundamental research questions and discuss existing solutions, open challenges, and research opportunities in B5G and 6G networks.

This article investigates the deployment of a Near-Real-Time Radio Access Network (RAN) Intelligent Controller (near-RT RIC) on a distributed cloud infrastructure composed of multiple physical sites with different amounts of resources and associated costs. The challenge is dynamically adapting the near-RT RIC deployment to the most cost-effective arrangement while meeting the latency requirements between the near-RT RIC and the controlled nodes. We introduce an optimization model to solve the disaggregated near-RT RIC placement problem, considering a cloud-native infrastructure to minimize the placement cost while satisfying the latency-sensitive control loop requirements across the cloud-edge continuum. Moreover, we describe an experimental environment we created using geographically disparate cloud sites. We present data detailing the latencies of the communication links among these sites and the costs incurred in using this real-world infrastructure. We conduct a performance evaluation of the near-RT RIC deployment, comparing the distributed approach versus a traditional monolithic strategy and evaluating positioning costs, deployment, setup and registration times, and the control loop latency considering three scenarios. Our results show that in a cloud-native environment, the disaggregated near-RT RIC allows cost savings of up to 60% in comparison to a monolithic near-RT RIC while satisfying the control loop latency and achieving time efficiency in terms of deployment and registration of xApps and near-RT RIC components.

Network slicing (NS) is becoming an essential element of service management and orchestration in communication networks, starting from mobile cellular networks and extending to a global initiative. NS can reshape the deployment and operation of traditional services, support the introduction of new ones, vastly advance how resource allocation performs in networks, and notably change the user experience. Most of these promises still need to reach the real world, but they have already demonstrated their capabilities in many experimental infrastructures. However, complexity, scale, and dynamism are pressuring for a Machine Learning (ML)-enabled NS approach in which autonomy and efficiency are critical features. This trend is relatively new but growing fast and attracting much attention. This article surveys Artificial Intelligence-enabled NS and its potential use in current and future infrastructures. We have covered state-of-the-art ML-enabled NS for all network segments and organized the literature according to the phases of the NS life cycle. We also discuss challenges and opportunities in research on this topic.

The fifth-generation mobile evolution introduces Next-Generation Radio Access Networks (NG-RAN), splitting the RAN protocol stack into the eight disaggregated options combined into three network units, i.e., Central, Distributed, and Radio. The disaggregated units reach full interoperability on Open RAN. Further advances allow the RAN software to be virtualized (vNG-RAN) on top of general-purpose hardware, enabling the management of disaggregated units and protocols as radio functions. The placement of these functions is challenging since the best decision must be based on multiple constraints, e.g., the RAN protocol stack split, routing paths in network topologies with restricted bandwidth and latency, asymmetric computational resources, etc. The literature does not deal with general placement problems with high functional split options and protocol stack analysis. This article proposes the first exact model for positioning radio functions for vNG-RAN planning, named PlaceRAN, as a Binary Integer Linear Programming (BILP) problem. The objective is to minimize the computing resources and maximize the aggregation of radio functions. The evaluation considered two realistic network topologies, and the results reveal that PlaceRAN achieves an optimized high-performance aggregation level. It is flexible for RAN deployment overcoming the network restrictions, and up to date with the most advanced vNG-RAN design and development.

The fifth-generation mobile evolution enables Next-Generation Radio Access Networks (NG-RAN) transformations. The RAN protocol stack is split into eight disaggregated options combined in three network units, i.e., Central, Distributed, and Radio. Further advances allow the RAN functions to be virtualized on top of general-purpose hardware using the virtualized RAN (vRAN). The combination of NG-RAN and vRAN results in vNG-RAN, enabling the management of the disaggregated units and protocols as a set of radio functions. However, the orchestration-based placement of these radio functions is challenging since the best decision can be determined by multiple constraints involving RAN disaggregation, crosshaul network requirements, availability of computational resources, etc. This article proposes OPlaceRAN, a vNG-RAN deployment orchestrator framed within the NFV reference architecture and aligned with the Open RAN initiative. OPlaceRAN supports the dynamic placement of radio functions focusing on vNG-RAN planning and is designed to be agnostic to the placement optimization solution. We developed a prototype based on cloud-native tools to deploy RAN using containerized virtualization and the OpenAirInterface emulator. The evaluation is analyzed considering two different approaches as a proof-of-concept. First, we applied two placement solutions in a controlled real computing infrastructure with a crosshaul network. Second, we investigated the orchestrator’s scalability with a real and larger-scale topology. Our results show that OPlaceRAN is an effective cloud-native solution for containerized network function placement and agnostic to the placement solution, handling scale-out well. OPlaceRAN is up-to-date with the most advanced vNG-RAN design and development approaches, contributing to the evolution of fifth-generation networks.

Unmanned Aerial Vehicles (UAVs) are being developed and researched to be used in different areas and diverse use cases. The delivery of medical parcels is an emergency service that avails from the air traffic for the fast delivery of critical health content. Besides 5G networks being the target communication technology for the UAV parcel transport, the infrastructure is still not optimized for ground-level User Equipments (UEs). Using multiple Radio Access Technology (RAT) by UAVs is an alternative solution for scenarios where the 5G network is unstable or not yet fully deployed. In this paper, we present bench-marking of 5G and 4G networks on unideal radio scenarios to understand the real latency and throughput delivered by these technologies in worst-case scenarios. The results show that the 5G network can provide the necessary Quality of Service (QoS) for UAV operations even with Reference Signal Received Power (RSRP) in the order of -90. In comparison, the preliminary data for the 4G network validates that the 4G network could not provide the required latency.

Openness, virtualization and disaggregation of functions represent the state-of-the-art (SOTA) for optimal management and orchestration of Radio Access Network (RAN) resources. However, in 5G and beyond networks, virtualized RAN (vRAN) functions may commonly share computing resources with Multi-access Edge Computing (MEC) services. This paper introduces a new problem formulation that jointly optimizes vRAN functions and MEC services respecting the maximum acceptable delay of the applications. We show that our model achieves better solutions than a SOTA approach and it is also more flexible. We also present a heuristic solution that is able to achieve near optimal results for real-world networks.

Technological advances in the fifth-generation (5G) mobile networks are based on native cloud computing platforms, and Kubernetes (K8S) has emerged as a relevant orchestration system in this context. However, these platforms were not designed to support 5G services natively. To illustrate, Kubernetes is designed to be agnostic to the services that it orchestrates and cannot dynamically reconfigure the 5G core according to existing network resources, i.e., it provides a partial dynamic orchestration to perform network slicing. This paper proposes a solution integrated with K8S to allow full dynamic orchestration of network slicing at runtime, including online adjustments in the 5G core. This integration is accomplished through a K8S-integrated controller for the control plane. The controller adjusts the 5G core and adapts the virtualized infrastructure. The results show a reconfiguration based on full dynamic orchestration without interruption of the services provided, reducing by close to 50% the full reconfiguration requests number by network slices.

The converging ecosystem provided by Multi-access Edge Computing (MEC) has motivated novel DASH video streaming provisioning scenarios involving the joint coordination of different mechanisms for caching, communication and control. Given the complexity of designing such mechanisms, it is important to provide the research community with open-source tools that support the assessment of their feasibility, specially in real-world environment. Current network emulators still require a significant programming effort to meeting this need. To fill this gap, a new DASH emulator called QoE-DASH is presented in this work. QoE-DASH builds upon goDASH to evaluate the QoE of users consuming DASH content, taking into account network properties, user preferences, and context information. To demonstrate the capabilities of QoE-DASH, we exercise different functionalities of our tool and present a case study where two joint caching and recommendation models, proposed in the literature, are evaluated and their effects on user QoE are depicted using state-of-the-art QoE metrics.

While energy efficiency is an important issue in virtualized RAN due to its impact on OPEX, the centralization level of virtualized RAN functions is another relevant concern that can conflict with the former. In this paper, we introduce a bi-objective problem formulation representing these two objectives and solution strategy based on the ϵ-constraint approach to generate the minimal complete set of Pareto-optimal solutions. We investigate the trade-off between energy efficiency and centralization level in traditional and next-generation RAN topologies. We show scenarios allowing noticeable improvement in the centralization level (e.g., from near 10% to 30%) without impacting energy consumption. However, after a certain value of centralization level, the impact in the energy consumption may become high and hard to justify.

Future mobile networks are envisioned to become multi-service systems, enabling the dynamic deployment of services with vastly different performance requirements, accommodating the needs of diverse service providers. Virtualizing the mobile network infrastructure is of fundamental importance for realizing this vision in a cost-effective manner. While there have been extensive research efforts in virtualization for the mobile core network, virtualization in the radio access network (RAN) is still at an early stage. In this article, we present AIRTIME, a new RAN slicing system that enables the dynamic on-the-fly virtualization of RANs, with the programmability required by service providers to customize any aspect of their virtual RAN to meet their service needs. We present a prototype implementation of AIRTIME and evaluate the: (i) capacity to create virtual RANs on-the-fly, (ii) performance experienced by slice owners, (iii) isolation among multiple virtual RANs sharing the same physical infrastructure, and (iv) scalability to accommodate a large number of virtual RANs.

Virtual Reality (VR) content is gaining popularity and allowing users to immerse themselves in a new world over the Internet. However, the high-demand for resources and the low latency requirements of VR services require changes in the current 5G networks to deliver VR with quality assurance. Microservices present a suitable model for deploying services at different levels of a 5G fog computing architecture for managing traffic and providing Quality of Experience (QoE) guarantees to VR clients. However, finding the most suitable fog node to allocate microservices for VR clients in QoE-aware 5G scenarios is a difficult task. This article proposes a QoE VR-based mechanism for allocating microservice dynamically in 5G architectures, called Fog4VR. Fog4VR determines the optimal fog node to allocate the VR microservice based on delay, migration time, and resource utilization rate. This article also presents the INFORMER, an integer linear programming model aiming to find the optimal global solution for microservice allocation. Results obtained with INFORMER serve as a baseline to evaluate Fog4VR in different scenarios using a simulation environment. Results demonstrate the efficiency of Fog4VR compared to existing mechanisms in terms of cost, migration time, fairness index, and QoE.

Abnormal behavior in mobile cellular networks can cause network faults and consequent cell outages, a major reason for operational cost increase and revenue loss for operators. Nonetheless, network faults and cell outages can be avoided by monitoring abnormal situations in the network and acting accordingly. Thus, anomaly detection is an important component of self-healing control and network management. Network operators may use the detected abnormal behavior to quantify numerically their intensity. The quantification of abnormal behavior assists the characterization of potential regions for infrastructure updates and to support the creation of public policies for local connectivity enhancements. We propose an unsupervised learning solution for anomaly detection in mobile networks using Call Detail Records (CDR) data. We evaluate our solution using a real CDR data set provided by an Italian operator and compare it against other state-of-the-art solutions, showing a performance improvement of around 35%. We also demonstrate the relevance of considering the distinct traffic patterns of diverging geographic areas for anomaly detection in mobile networks, an aspect often ignored in the literature.

The computer network industry is preparing for the revolution caused by the virtualization of the network infrastructure. The virtualization of network functions, together with the network programmability, is a fundamental skill-set required in both electrical engineers and computer scientists, which will become even more significant in the upcoming years. Therefore, in this article, we present an educational framework for Service Function Chaining (SFC) practical teaching to undergraduate students aiming to prepare them for future Information and Communication Technologies (ICT) and communication networks market that will demand skillful professionals in the domain. The educational framework was designed for the Network Management course at the University of Antwerp. To structure the content of the framework’s sessions, we explain the fundamental concepts behind the SFC. Moreover, we detail the framework objectives, sessions, and how we plan to evaluate it. We also present facts and works that endorse the hands-on teaching preparing students for the research and development in the new generation communication networks arena. After, we discuss the methodology of the educational framework alignment with joint ACM/IEEE Computer Engineering Curricula for Network Management courses. Next, we present possible outcomes during the course evaluation using the survey. We conclude which this educational framework achieves the necessary core learning outcomes to empower students concerning the practical steps when deploying and managing communication networks and the virtualization techniques which nowadays are being applied.

In this work, we investigate opportunities for network operators to reduce their expenditures through optimal allocation of virtual Base-Band Units (vBBUs) in Fog Radio Access Networks (F-RANs). The optimal allocation can generate additional revenue opportunities by leasing idle processing resources to Application Service Providers (ASPs). In particular, we address the challenge of improving vBBU allocation in terms of optimal assignment of the workloads of Remote Radio Heads (RRHs) to Micro Data Centers (MDCs) for cost minimisation, considering the trade-off between MDC and RRH distance and processing power consumption. Thus, we propose an optimisation model to decide the assignments between MDCs to RRHs. The optimal solution is obtained through Binary Integer Linear Programming (BILP). We evaluate our solution by applying a real Call Detail Record (CDR) data set, simulating different regions from Milan. K-means clustering was used to identify the Internet traffic behaviour of different regions in Milan. This work results' highlight opportunities for network operators to exploit their infrastructure usage and increase their gains.

Long Range Wide Area Network (LoRaWAN) enables flexible long-range communication with low power consumption and low-cost design perspectives. However, the adoption of this technology brings new challenges due to the densification of IoT devices, which causes signal interference and affects the QoS directly. On the other hand, the flexibility in the LoRaWAN transmission configurations allows higher management in the use of end-device parameters, which allows better resource utilization and improves network scalability. This paper proposes an adaptive solution to handle the define best LoRaWAN parameter settings to reduce the channel utilization and, consequently, maximize the number of packets delivered. Additionally, to validate our method, we formulated mixed-Integer linear programming and results compared to those given by the heuristics. Results provided by the heuristic are close to those provided by the MILP.

Video-on-Demand (VoD) services create a demand for content orchestrator mechanisms to support Quality of Experience (QoE). Fog computing brings benefits for enhancing the QoE for VoD services by caching the content closer to the user in a multi-tier fog architecture, considering their available resources to improve QoE. In this context, it is mandatory to consider network, fog node, and user metrics to choose an appropriate fog node to distribute videos with QoE support properly. In this article, we introduce a content orchestrator mechanism, called of Fog4Video, which chooses an appropriate fog node to download video content. The mechanism considers the available bandwidth, delay, and cost, besides the QoE metrics for VoD, namely number of stalls and stalls duration, to deploy VoD services in the opportune fog node. Decision-making acknowledges periodical reports of QoE from the clients to assess the video streaming from each fog node. These values serve as inputs for a real-time Analytic Hierarchy Process method to compute the influence factor for each parameter and compute the QoE improvement potential of the fog node. Fog4Video is executed in fog nodes organized in multiple tiers, having different characteristics to provide VoD services. Simulation results demonstrate that Fog4Video transmits adapted videos with 30% higher QoE and reduced monetary cost up to 24% than other content request mechanisms.

Dynamic Cloud Radio Access Networks (C-RAN) is an emerging wireless architecture that splits radio functions in a hierarchical cloud-based infrastructure. In Dynamic C-RAN, wireless functionalities can be divided into smaller components and distributed along with a hierarchical cloud infrastructure. Network Functions Virtualization (NFV) concepts have been recently investigated to facilitate management-related operations of these wireless functionalities. Despite the advances in providing flexible NFV orchestration for Virtualized Network Functions (VNF)s, there is still a lack of solutions aware of VNFs internal composition, a key characteristic of split virtualized radio functions. In this article, we advance NFV orchestration in a Dynamic C-RAN scenario presenting Orchestra, a split-aware NFVO for Dynamic C-RAN with support to customizable orchestration algorithms for different radio function splits. Orchestra is validated through instantiates and migrates VNFCs following the decisions of personalized orchestration algorithms, with a petty impact in the VNFC's deployment and migration time.

Quality of Experience (QoE) should be the driver for network orchestration in 4G networks. At the same time, the network must be able to cope with high bandwidth requirements from applications such as video streaming, while dealing with a large number of users. This paper proposes a network orchestrator that adjusts network parameters to improve QoE of video streaming. The orchestrator uses Device-to-Device (D2D) communication to improve user's QoE, also reducing the demand on 4G network. The use of D2D is triggered by a machine learning engine. Experiments made in a physical testbed show an improvement on the mean horizontal video resolution from 768 to 1280 pixels, as well as a decrease of around 90% at the impact on the QoE, considering the number of video resolution changes. Finally, the demand on the network backhaul is decreased by around 38%.

The fifth generation of mobile networks is envisioned to provide connectivity services to a multitude of devices with vastly different requirements. Current mobile systems rely on inflexible hardware- based RF front- end that provide a "one-size-fits-all" air-interface. Instead, future mobile networks should be flexible, providing different air-interfaces for particular users and applications. In this paper, we present HyDRA, a software-defined radio virtualization layer that enables the execution of multiple programmable air-interfaces on top of one RF front-end. Our solution multiplexes digitized IQ signal samples of multiple virtual radios into a single stream. We have implemented HyDRA and experimentally evaluate its performance in a scenario that considers a base station executing LTE and NB-IoT VRs. Results obtained show that HyDRA is able to efficiently multiplex these two technologies, while the computational analysis shows that HyDRA is not CPU-intensive and can run in standard, commodity computers. HyDRA is a promising framework to enable RRH slicing, multi-radio access networks, and flexible multi-tenant networks.

Dynamic Cloud Radio Access Network (Dynamic C-RAN) is an emerging wireless architecture that aims for flexibility, business agility, adaptability, among other benefits. In a Dynamic C-RAN, wireless functionalities can be split into smaller components and distributed along a hierarchical cloud infrastructure. Network Functions Virtualization (NFV) concepts have been recently investigated to facilitate management-related operations of these wireless functionalities. Despite the many advocated advantages of the function's splitting and the effectiveness of NFV orchestration solutions, both academia and industry are considering Virtualized Network Functions (VNF)s as atomic elements, disregarding the potential advantages of splitting VNFs into several different components. Aiming to improve VNF orchestration in Dynamic C-RAN scenarios, in this paper we propose Maestro: an NFV orchestrator for wireless environments that is able to decide among several possible VNF compositions which are more suitable for each situation. Maestro is designed to operate using different decision mechanisms that can be defined based on network operators' needs. We evaluate the effectiveness of our proposal by modeling the orchestrator's decision mechanism as a linear programming problem. Thus, we show how fronthaul bandwidth consumption can be reduced threefold considering different VNF compositions against atomic VNF placement.

Mobile Health broadens the accessibility to healthcare applications using mobile devices, such as smartphones. These devices can be used for tracking the patient health signals for remote monitoring and treatment follow up. Sleep disorders affect a substantial part of the population and may have serious comorbidities. To improve monitoring of evolution of sleep disorders and/or treatment follow up, we propose an unsupervised segmentation and classification of snoring events (as snore or non-snore) on a Mobile Health context. We apply a statistical analysis through the Expectation- Maximization algorithm applied to Gaussian Mixture Models to cluster the data sets. We evaluate our solution using recordings from simple and Obstructive Sleep Apnea (OSA) snorers. The results show that our proposal is able to classify snoring events achieving average accuracies of 91.3% for simple snorers and 79.7% for OSA snorers.

We propose an Adaptive Monte Carlo algorithm to perform global energy efficient resource allocation for Heterogeneous Cloud Radio Access Network (H-CRAN) architectures, which are forecast as future fifth-generation (5G) networks. We argue that our global proposal offers an efficient solution to the resource allocation for both high and low density scenarios. Our contributions are threefold: (i) the proposal of a global approach to the radio resource assignment problem in H-CRAN architecture, whose stochastic character ensures an overall solution space sampling; (ii) a critical comparison between our global solution and a local model; (iii) the demonstration that for high density scenarios Energy Efficiency is not a well suited metric for efficient allocation, considering data rate capacity, fairness, and served users. Moreover, we compare our proposal against three state-of-the-art resource allocation algorithms for 5G networks.

Software-Defined Networking (SDN) is an emerging paradigm that arguably facilitates network innovation and simplifies network management. However, in the context of SDN, management activities, such as monitoring, visualization, and configuration can be considerably different from traditional networks. An SDN controller, for example, can be customized by network administrators according to their needs. Such customizations might pose an impact on resource consumption and traffic forwarding performance, which is difficult to assess without an SDN-devoted management system. In this paper, we initially present an analysis of control traffic in SDN aiming to better understand the impact of the communication between the controller and forwarding devices. Afterwards, we propose an interactive approach to SDN management through monitoring, visualization, and configuration that includes the administrator in the management loop. To show the feasibility of our approach a prototype has been developed. The results obtained with this prototype show that our approach can help the administrator to better understand the impact of configuring SDN-related parameters on the overall network performance.

Currently, many investigations addressed SDN management considering using monitoring information for different purposes. Zhang [1] used monitoring information to develop algorithms for anomaly detection and traffic engineering. Jose et al. [2] used the same monitoring information to propose mechanisms for online measurement of large traffic aggregates. Yu et al. [3] proposed FlowSense, which is aimed to keep the lowest control channel overhead and the highest information accuracy as possible in OpenFlow monitoring. Chowdhury et al. addressed SDN monitoring from a different perspective. The authors proposed Payless [4] framework that is able to deal with monitoring considering the polling frequency and data granularity. Payless relies on OpenTM [5] to select only important switches to be monitored, also aiming to reduce the overhead imposed in the control channel. Thus, this framework allows to adjust the polling frequency parameter to balance the control channel overhead, imposed by monitoring messages, and accuracy of monitored information.

Cognitive radio make use of spectrum sensing techniques to detect licensed users transmissions and avoid causing interference. The major drawback in current spectrum sensing techniques is the use of static decision thresholds to detect such transmissions, which may be infeasible in public safety radio channels. More precisely, the cognitive radio may find different noise or interference levels when switching among these channels. This can lead to a wrong picture of the channel occupancy status, which in turn can increase the interference caused to licensed users. In this paper we propose an Adaptive Threshold Architecture, which uses machine learning algorithms to dynamically adapt the decision threshold, enabling the detection of licensed users transmissions in public safety radio channels. Results showed that the proposed architecture increased the sensing accuracy up to 2 times, providing results up to 6 times faster when compared to other solutions of the literature.

Software defined radio enables the improvement of the radio-frequency spectrum utilization through the design of cognitive radio devices. The implementation of these devices must be based on spectrum sensing function searching for vacant channels and, opportunistically, transmit over these channels in a cognitive radio network. Therefore, the configuration, monitoring and visualization of the spectrum sensing function are fundamentals to the continuous learning process of the network administrator. In this paper we propose Kitsune, a management system based on a hierarchical model allowing to manage summarized information about the spectrum sensing function in a cognitive radio networks. Moreover, a Kitsune prototype was developed and evaluated through a real IEEE 802.22 scenario using TV channels to Internet access. Results shown that Kitsune allows network administrator to achieve a higher knowledge about behavior of the users and improve the average throughput for each channel.

Cognitive Radio Networks enable a higher number of users to access the spectrum of frequency simultaneously. This access is possible due to the implementation of dynamic spectrum allocation algorithms. In this context, one of the main algorithms found in the literature is the reinforcement learning based approach called Q-Learning. Although been widely applied, this algorithm does not take into account accurate information about the behavior of users neither the channel propagation conditions. In this sense, we propose three improvements to the dynamic spectrum allocation algorithms based on reinforcement learning for cognitive sensor networks. Simulation results show that all the proposed algorithms allow allocating channels with up to 6dB better quality and 4% higher efficiency than Q-Learning.

The main goal of change management is to ensure that standardized methods and procedures are used for the efficient and prompt handling of changes in IT systems, in order to minimize change-related incidents and service-delivery disruption. To meet this goal, it is of paramount importance reusing the experience acquired from previous changes in the design of subsequent ones. Two distinct approaches may be usefully combined to this end. In a top-down approach, IT operators may manually design change templates based on the knowledge owned/acquired in the past. Considering a reverse, bottom-up perspective, these templates could be discovered from past execution traces gathered from IT provisioning tools. While the former has been satisfactorily explored in previous investigations, the latter - despite its undeniable potential to result in accurate templates in a reduced time scale - has not been subject of research, as far as the authors are aware of, by the service operations and management community. To fill in this gap, this paper proposes a solution, inspired on process mining techniques, to discover change templates from past changes. The solution is analyzed through a prototypical implementation of a change template miner subsystem called CHANGEMINER, and a set of experiments based on a real-life scenario.

In order to deal with failures in the deployment of IT changes and to always leave IT infrastructures into consistent states, we proposed in a previous work, a solution to automate the generation of rollback plans in IT change management systems. The solution was based on a mechanism that treats Requests for Change (RFC) (or parts of them) as a single atomic transaction. In this work, we extend our previous investigation and present more flexible and fine grained treatment of failures. The paper first presents extensions to our conceptual model in order (i) to give IT operators some flexibility in defining rollback actions, for example, by allowing the rollback plan to not only be a reversed change plan; and (ii) to execute different recovery activities depending on the cause and location of a problem. The paper then focuses on a refined manner to handle and treat failures in change deployments. We follow the ITIL version 3 best practises which suggest that, depending on the RFC context, the human operator can classify activities as reversible or irreversible. Such classification allows change management systems to automatically generate more accurate remediation plans. The proposal takes into account not only a precise way to define how rollback plans will be generated, but also an intuitive method enabling the operator to define compensation activities in order to complete the RFC successfully, even with the occurrence of failures. To prove the concept and technical feasibility, we have materialized our solution in the CHANGELEDGE prototype that, using elements of the Business Process Execution Language (BPEL), is able to generate correct remediation plans to handle and treat failures in IT change management systems.

The growing necessity of organizations in using technologies to support to their operations implies that managing IT resources became a mission-critical issue for the health of the primary companies' businesses. Thus, in order to minimize problems in the IT infrastructure, possibly affecting the daily business operations, risks intrinsic to the change process have to be analyzed and assessed. Risk Management is a widely discussed subject in several areas, although for IT Change Management it is quite a new discipline. The Information Technology Infrastructure Library (ITIL) introduces a set of best practices to conduct the management of IT infrastructures. According to ITIL, risks should be investigated, measured, and mitigated before any change is approved. Even with these guidelines, there is no default automatic method for risk assessment in IT Change Management. In this paper we introduce a risk analysis method based on the execution history of past changes. In addition, we propose a failure representation model to capture the feedback of the execution of changes over IT infrastructures.

Capturing and reusing the experience of operators in implementing IT changes is an important aspect of IT service management, as it may result in fewer incidents (upon change execution) and faster specification of change plans, to mention just a few potential advantages. Nevertheless, in practice, changes are usually described and documented in an ad hoc fashion, due to the lack of proper support to assist the design process. This hampers knowledge acquired when specifying, planning, and carrying out previous changes to be reused in subsequent requests. In order to address this issue, we propose the use of change templates as a mechanism to formalize, preserve, and reuse the experience accumulated within organizations in relation to IT changes. Our solution is analyzed through a prototypical implementation of a change management system and a case study based on a real-life scenario.

The current research on IT change management has been exploring several aspects of this new discipline, but it usually assumes that changes expressed in requests for change (RFC) documents will be successfully executed over the managed IT infrastructure. This assumption, however, is not realistic in actual IT systems because failures during the execution of changes do happen and cannot be ignored. In order to address this issue, we propose a solution where tightly-related change activities are grouped together forming atomic groups of activities. These groups are atomic in the sense that if one activity fails, all other already executed activities of the same group must rollback to move the system backwards to the previous state. The automation of change rollback is especially convenient because it relieves the IT human operator of manually undoing the activities of a change group that has failed. To prove concept and technical feasibility, we have materialized our solution in a prototype system that, using elements of the business process execution language (BPEL), is able to control how atomic groups of activities must be handled in IT change management systems.