Fault detection is a critical problem in Open Radio Access Network (O-RAN) observability, but the availability of labeled fault data is extremely limited, and enumerating all possible fault scenarios is costly and often infeasible. Therefore, it is essential to rely on normal network behavior to enable effective fault detection. In this work, we propose FALCON++, a tailored semi-supervised anomaly detection and localization framework. Specifically, we use a specialized contextual autoencoder that fuses feature-wise gating and hierarchical temporal encoding for the effective reconstruction of normal network behavior. Furthermore, we combine a causal decoder with residual skip connections to improve robustness to data deemed anomalous due to their higher reconstruction error. FALCON++ also addresses scalability issues by shifting from a monolithic system model to CU-DU pairwise inferencing. Through experimental results on synthetic faults generated using STRESS-NG and TC, we show that FALCON++ achieves a high-precision and high-recall operating point with an F1-score of 0.77, thereby enabling anomaly detection and localization.

Emerging network technologies have the potential to significantly reduce End-to-End (E2E) service delays for applications with stringent quality of service (QoS) requirements. Network Function Virtualization (NFV) enables dynamic orchestration of Service Function Chains (SFCs), allowing client data to sequentially traverse multiple Virtual Network Functions (VNFs) for processing and analysis. However, this sequential processing introduces additional delays. To address this, VNF Parallelization (VNFP) has been proposed, enabling multiple VNFs to run parallel, thereby significantly reducing SFC E2E delay. Despite its benefits, VNFP introduces two key challenges: (i) packet copy/merge overheads, which involve creating duplicate packets for parallel VNFs, and (ii) packet deposition overheads, which occur due to variations in VNF processing times, necessitating storage of early-processed packets until all packets can be merged. Existing approaches have overlooked these overheads during VNF selection. In this paper, we address these overheads for the first time by considering them in the VNF instance selection process while ensuring that SFCs meet the E2E delay requirements. We formulate the VNFs selection problem as an Integer Linear Program (ILP) that minimizes VNFP overheads by jointly selecting VNFs and routing traffic for each request from source to destination. Given the NP-hard nature of this problem, we propose a heuristic algorithm named Parallelism Diligent Algorithm (PDA), whose PSFC-structuring step is based on a first-fit decreasing bin-packing procedure with a known approximation bound. Our simulation results demonstrate that PDA achieves performance close to the ILP optimum in the small-scale instances where the ILP can be solved, and outperforms the considered benchmark schemes in the evaluated scenarios.

Integrated Sensing and Communications (ISAC) enhances traditional mobile network capabilities by enabling the detection of passive, non-connected objects. Latency-sensitive vehicular applications such as Augmented Reality (AR), Virtual Reality (VR), and High Definition Maps (HD Maps) can be integrated with ISAC to improve the utilization of limited wireless resources. The Configuration Grant (CG) mechanism, in 3GPP Release 16, reduces signaling overhead in Uplink (UL) by preassigning resources to UEs (vehicles). However, employing CG for ISAC can lead to incorrect assignment of transmission slots due to the aperiodic nature of ISAC sensing traffic. To address this issue, we propose a CG allocation scheme that models the interarrival times of aperiodic traffic using a Weibull distribution. A probability distribution model, implemented and evaluated using the NS-3 5G-LENA CG module, assists the radio resource scheduler by analyzing sensing arrivals within a configuration window to predict future bursts and proactively reserve UL resources for UEs (vehicles). This prediction-driven allocation significantly improves Packet Delivery Ratio (PDR) and spectral efficiency in vehicular scenarios.

In dynamic hotspot scenarios, unserved users from the hotspots of the overloaded gNBs can be offloaded to the neighbouring underloaded gNBs by using Unmanned Aerial Vehicles (UAVs) as relays. However, it may lead to co-channel interference as the same frequencies are used by every gNB. Therefore, unlicensed bands are leveraged for communication between the offloaded users and the UAVs in the state-of-the-art. It leads to contention with other users that use unlicensed bands and reduces the efficiency of spectral usage. To avoid the usage of unlicensed bands, we propose an offloading solution where Successive Interference Cancellation (SIC) is used to handle the co-channel interference. The proposed solution can be used for both cell-centred and cell-edge hotspot scenarios. New challenges arise when a multi-gNB and multi-hotspot scenario is considered, such as determining the load re-distribution strategy, determining the number and locations of the UAVs, detecting the hotspots across the coverage area of multiple gNBs, and determining the user offloading strategy that makes SIC feasible. Therefore, we formulate the problem for multi-gNB and multi-hotspot scenario to minimize the total cost of Service Level Agreement (SLA) violations, by jointly optimizing the load redistribution, number and locations of the UAVs, power levels of the gNBs and UAVs, bandwidth allocation and user offloading strategy. We also propose an efficient heuristic solution to the formulated problem and demonstrate its effectiveness through extensive numerical simulations. The numerical simulations show that the proposed solution achieves up to 55.39% fewer SLA violations and up to 37.91% less total SLA violation cost compared to the baseline.

Emerging latency-critical applications such as cloud gaming and industrial automation demand agile and ultra-low-latency service delivery, which traditional network appliances struggle to support. Network Function Virtualization (NFV) addresses this by chaining Virtual Network Functions (VNFs) into Service Function Chains (SFCs). Parallelized SFCs (PSFCs) reduce service delay by executing independent VNFs concurrently, but introduce significant copy/merge and buffering overheads due to synchronization delays across branches. Moreover, dynamic PSFC arrival rates complicate efficient VNF placement decisions. This paper presents RL-PARETO, a hybrid deep reinforcement learning approach that adaptively orchestrates parallel VNFs while minimizing parallelization overheads and satisfying SLA constraints. RL-PARETO uses a graph transformer encoder with dual pointer-network heads to jointly generate PSFC partitions and VNF placements in a single pass. Training integrates Proximal Policy Optimization (PPO) for stable exploration with a Double-DQN critic for efficient value estimation. A fallback heuristic ensures feasible deployments under resource constraints. Extensive evaluations across diverse network topologies demonstrate that RL-PARETO achieves up to $10 \%$ higher acceptance rate and $15 \%$ reduction in merge buffer overhead, while maintaining robust performance under dynamic conditions.

In this paper, we address the problem of reconfiguration of network slices (NSs) to handle network node failures and mitigate service interruptions while satisfying the diverse service requirements of users and multiple industry verticals in beyond 5G and 6G networks. First, we formulate the latency- and reliability-aware NS reconfiguration problem as an integer linear programming (ILP) to minimize resource consumption. Then, to deal with the computational time complexity of ILP, we propose a matching algorithm-based heuristic solution. Through extensive simulations, we show that our proposed algorithm provides a nearoptimal solution in polynomial time with an optimality gap of 2.5% and consumes fewer resources while satisfying stringent latency and reliability service requirements.

Network Functions Virtualization (NFV) leverages virtualization and cloud computing technologies to make networks more flexible, manageable, and scalable. Instead of using traditional hardware middleboxes, NFV uses more flexible Virtual Network Functions (VNFs) running on commodity servers. One of the key challenges in NFV is to ensure strict reliability and low latency while also improving energy efficiency. Any software or hardware failures in an NFV environment can disrupt the service provided by a chain of VNFs, known as a Service Function Chain (SFC), resulting in significant data loss, delays, and wasted resources. Due to the sequential nature of SFC, latency increases linearly with the number of VNFs. To address this issue, researchers have proposed parallelized SFC or VNF parallelization, which allows multiple independent VNFs in an SFC to run in parallel. In this work, we propose a method to solve the parallelized SFC deployment problem as an Integer Linear Program (ILP) that minimizes energy consumption while ensuring reliability and delay constraints. Since the problem is NP-hard, we also propose a heuristic scheme named ERASE that determines the placement of VNFs and routes traffic through them in a way that minimizes energy consumption while meeting capacity, reliability, and delay requirements. The effectiveness of ERASE is evaluated through extensive simulations and it is shown to perform better than benchmark schemes in terms of total energy consumption and reliability achieved.

Network Functions Virtualization (NFV) replaces traditional hardware-based network equipment and middleboxes with flexible Virtualized Network Functions (VNFs) in order to reduce costs and improve agility and scalability. The VNFs are logically arranged in a specific sequence to form a Service Function Chain (SFC) which ensures that the traffic is processed according to the desired service requirements. However, the inherent length of SFCs leads to an undesirable increase in end-to-end delay experienced by the packets. Parallelized SFC (PSFC) addresses this problem by trying to allow multiple VNFs of the SFC to process packets in parallel by co-locating parallelizable VNFs on the same server. The energy-efficient deployment of PSFCs while considering the impact of contention for the shared resources on the server is unexplored in the literature. Hence, in this work, we formulate the PSFC deployment problem as an Integer Linear Program (ILP) that minimizes energy consumption while considering the impact of shared resource contentions without violating end-to-delay constraints. Since the ILP is NP-hard, we also propose a heuristic scheme named EPSFC, which provides flexible resource allocation-based deployment that minimizes the total energy consumption and ensures end-to-end delay requirements while considering the effects of shared resource contentions on the end-to-end delay. The effectiveness of EPSFC is evaluated through extensive simulations, and the results show a significant reduction in energy consumption while improving the PSFC acceptance ratio as compared to state-of-the-art schemes.

New Radio (NR) Sidelink was first introduced by 3GPP in Rel. 16 to meet service requirements of advanced Vehicle-to-Everything (V2X) use cases. The mode-2 of NR-V2X enables direct vehicle-to-vehicle communication through distributed resource scheduling, specifically Semi-Persistent Scheduling (SPS). The focus of Intelligent Transport Systems (ITS) till now has been predominantly on ground vehicles and infrastructure, however the anticipated future scenario where airspace becomes as crowded as ground space by the mid-2030s necessitates a robust and efficient cooperation between aerial vehicles (AVs) and terrestrial/ground vehicles (GVs). There is also an ongoing discussion in the 3GPP Technical Specification Group Radio Access Network (TSG RAN) on supporting aerial communications in AVs by using NR-sidelink. This paper is the first to study extensively the performance of Cooperative Awareness Messages (CAM) and Decentralized Environmental Notification Messages (DENM) in a unified setup, involving systems of aerial vehicles (AVs) and systems of ground vehicles (GVs) in mode-2 using NR Sidelink. The SPS at Medium Access Control (MAC) layer schedules CAM and DENM packets. We conduct a comparative analysis of CAM performance for AVs versus GVs using a 3D geographical dataset of Washington DC, evaluating Packet Reception Ratio (PRR) and Packet Inter-Reception (PIR). Furthermore, we enhance the reachability of DENM among GVs by integrating AVs using the NR sidelink UE (User Equipment)-to-UE relay feature introduced in Rel. 17/18. Finally, we propose a congestion-aware multi-hop DENM dissemination protocol with a novel relay selection mechanism to keep congestion below the maximum Channel Busy Ratio (CBR) limit, thereby improving reception node coverage by 17% compared to state-of-the-art NR-V2X DENM dissemination mechanisms.

Enabling widespread adoption of resource-intensive vehicular applications such as Extended Reality (XR) and High Definition map (HD Map) necessitates further enhancements in 5G, which is anticipated with 5G-Advanced. These applications, sensitive to latency, prompt researchers to propose offloading vehicles' complex computations to nearby edge clouds, aiming to minimize latency and meeting the Quality-of-Service (QoS) demands of these applications. However, the uncertainties arising from spatio-temporal factors due to vehicle mobility and the dynamic nature of application behaviour pose significant challenges in deciding the efficient offloading decision for minimizing latency. To tackle this challenge, this paper introduces a crosslayer framework that bridges the Radio Access Network (RAN) scheduler with the Mobile Edge Computing (MEC) scheduler. The proposed framework facilitates the exchange of vehicle ranks and channel condition information between schedulers, strategically aimed at reducing Head-Of-Line (HOL) delay for efficient computational offloading. Furthermore, the MAC layer incorporates the prediction of the Buffer Status Report (BSR) using Machine Learning (ML) to further reduce the queuing delay experienced by the offloading jobs of the vehicles in uplink. Simulation results using the NS-3 gym demonstrate that the proposed cross-layer framework achieves a higher Offloading Success Rate (OSR) than the state-of-the-art QoS scheduler by effectively reducing HOL delay for HD Map vehicular application.

The private 5G-and-beyond networks have the ability to fulfill the high data rate, low latency, and massive connectivity requirements of industrial IoT (IIoT) automation. As a result, the slice selection requests arriving at the control plane functions (CPFs) of the private 5G-and-beyond networks will increase. The network has to scale up its resources to improve the response time for slice selection requests. The scaling up of resources will lead to an increase in operational expenditure (OPEX). We propose a novel approach to the slice selection method in private 5G-and-beyond networks for industrial IoT that improves response time for slice selection requests and reduces OPEX for the networks. The response time of our proposed method (protocol) for u IIoT devices is approximately 1/u times that of the traditional slice selection protocol, and the number of instances of CPFs required for the proposed protocol is also lesser than that of the slice selection protocol in 5Gs networks.

Cooperative Awareness Messages (CAMs) and Decentralized Environmental Notification Messages (DENMs) are two types of messages used in Intelligent Transport Systems (ITS) for Vehicle-to-Everything (V2X) communication. CAMs are used to periodically exchange the current state of a vehicle with its nearby vehicles and the DENMs are used to aperiodically provide critical and time-sensitive information about current environmental conditions to other vehicles and infrastructure entities. While considerable research has been conducted on the efficient scheduling of CAM and DENM separately or at the network level coexistence, no work has been done to integrate the scheduling of both message types at the granularity of the per-vehicle level. The scheduling of CAM is commonly done through Semi-Persistent Scheduling (SPS). However, no equivalent scheduling algorithm has been developed for the aperiodic and variable-sized DENMs. As a result, there is a need for novel mechanisms that can efficiently prioritize and schedule mixed traffic of CAM and DENM messages. This paper introduces a Quality of Service (QoS) scheduler called LAMP for segregating traffic in the Radio Link Control (RLC) layer for joint sidelink scheduling of CAM and DENM at the vehicular level. The LAMP scheduler is aided with a special resource selection and reservation scheme for mixed traffic scenarios. The groundwork involves an experimental analysis in Network Simulator-3 (NS-3) that uses the New Radio (NR) Vehicle-to-Everything (V2X) module in Mode-2. The simulation findings demonstrate that LAMP could significantly reduce the end-to-end latency of DENM by 89.36% and CAM by 40.2% while also increasing the packet reception rate by 12.1% and 9.74% for CAM and DENM with repetitions, respectively.

By bringing the processing and storage capabilities of the cloud closer to the end devices, fog computing (FC) enhances the Quality of Service (QoS) for latency-critical Internet of Things (IoT) applications, such as autonomous driving, haptics, and augmented reality (AR). To facilitate the processing and storage of data packets, the fog nodes in the underlying FC-enabled IoT network (FC-IoTN) are to be provisioned with storage and processing resources. Existing resource provisioning solutions focus mainly on latency sensitivity and cost efficiency. They also operate under the assumption that these fog nodes are completely reliable and energy efficient. In reality, this is not true. The fog nodes are not 100% reliable. Neither are they energy efficient. In this study, we propose a novel resource provisioning framework for the fog nodes that considers reliability and energy efficiency, in addition to latency sensitivity and cost efficiency. We first give an analytical framework to model the failures and recoveries in a fog node and use this modeling to provision resources in the fog nodes such that the resultant resource provisioning is optimal in terms of cost and energy consumption. Further, to understand the effect of latency, reliability, cost, and energy on resource provisioning, we analyze and decode the interplay between these factors during resource provisioning in fog nodes. We finally show the efficacy of our approach over the scenario that does not consider reliability and energy efficiency while provisioning resources. Without affecting the latency sensitivity and reliability of the system, our framework achieves an enhancement of 35%, and 37% in terms of cost and energy consumption, respectively, over a nonoptimized framework.

Vehicular applications such as Augmented Reality (AR), Virtual Reality (VR), and High Definition Map (HD Map) are known for their latency-sensitive traits. But, dynamic scheduling at the MAC layer incurs significant signalling overhead (in terms of Scheduling Requests (SRs) in Uplink (UL)), leading to non-negligible latency in 5G NR. To address this issue, 5G NR introduces Configuration Grant (CG) for UL transmission, which pre-allocates radio resources to UEs (vehicles), thereby reducing signalling overhead between a vehicle and the Base Station (gNB). However, the high-speed mobility of vehicles results in rapid changes in channel conditions. Employing CG in a vehicular scenario can lead to incorrect assignment of transmission parameters (e.g., Modulation and Coding Scheme (MCS)), thereby adversely impacting the vehicles’ Packet Delivery Ratio (PDR). To address this issue, this paper proposes a CG allocation algorithm that utilizes a Machine Learning (ML)-driven approach to predict the future MCS of vehicles. A data-driven ML model, derived from a real-world dataset, assists the radio resource scheduler and is evaluated using the NS-3 5G-LENA CG module. The ML-assisted CG allocation algorithm demonstrates significant improvements in terms of PDR and spectrum usage efficiency in vehicular scenarios.

Unmanned Aerial Vehicles (UAVs) can be leveraged to perform cellular user offloading in dynamic hotspot scenarios and enable the unserved users of overloaded gNodeB (gNB) use the unutilized bandwidth of a neighbouring underloaded gNB. To reach the blocked users and to minimize the co-channel interference between offloaded users and other users, we propose using UAVs mounted with two Intelligent Reflecting Surface (IRS) panels. We formulate the problem to minimize the total Service Level Agreement (SLA) violations cost by jointly optimizing the location of the UAV, user association, phase shifts of the IRS elements, transmit power allocation, and bandwidth allocation. Further, we propose an efficient heuristic solution to the formulated problem and show its effectiveness using numerical simulations. Through our simulations, we show that the total SLA violation cost decreases by up to 17.18% and spectral efficiency increases by up to 6.66% by using 400 elements in each IRS.

In this paper, we study an Unmanned Aerial Vehicle (UAV) enabled Mobile Edge Computing (MEC) service provisioning to the Internet of Remote Things (IoRT) devices spread randomly on the ground in a remote area. The data generated by the IoRT devices is collected by the UAVs, which immediately relay the data collected to an MEC device installed on the ground at a nearby location. The MEC device receives the data from the UAVs, and sends the results back to the UAVs, which in turn relay them to IoRT devices. We aim to minimize the energy consumption by the IoRT devices and the UAVs, while maximizing the system throughput subject to bandwidth, power, information-causality, and UAVs’ trajectory constraints. We formulate the problem as a Mixed Integer Non Linear Programming problem, which is a complex and non-convex optimization problem. To make the problem tractable, we use variable relaxation. We further develop an iterative algorithm based on Block Coordinate Descent method, to jointly optimize the connection scheduling, power control, bit transmission scheduling, bandwidth allocation, and trajectories of the UAVs. Numerical results demonstrate the convergence of the algorithm and superiority of the proposed model with respect to conventional methods. Our proposed system model of placing MEC at ground shows 9% improvement in energy consumption when compared to carrying out computations at MEC carried by UAV and a 99% improvement when compared to placing MEC at the satellite. The proposed system model shows a 0.2% lower system throughput on average, compared to placing MEC at UAV, which is tolerable considering gains in terms of energy consumption.

Using Unmanned Aerial Vehicles (UAVs) is a promising solution to perform cellular user offloading in dynamic hotspot scenarios. The users from overloaded gNodeB (gNB) can be offloaded to a neighbouring underloaded gNB using UAVs as relays. In this paper, we propose a novel UAV-aided cellular user offloading that uses Successive Interference Cancellation (SIC) for handling interference. Our proposed solution can be used for dynamic hotspots that occur in both cell-centered and cell-edge areas. We do not use any unlicensed band for communication between users and UAV, thereby avoid contention between cellular technologies and technologies that use unlicensed bands such as WiFi. As the users perform SIC for retrieving the signal, user association, location of UAV, and power levels of UAV and gNB are crucial for providing the promised data rates to the users, so that the Service Level Agreements (SLAs) are not violated. Therefore, we formulate the problem to minimize the SLA violations by jointly optimizing the location of UAV, power levels of gNB and UAV, bandwidth allocation and user association. Further, we propose a solution to the formulated problem and show the effectiveness of our proposed solution using extensive numerical simulations. The proposed solution is shown to significantly decrease the total cost of SLA violations and increase the total available rate for the users of the overloaded gNB.

The Internet of Things (IoT) is scaling rapidly to billions of low power devices, with diverse radio technologies sharing common unlicensed spectrum. Inevitably, this results in rampant cross-technology collisions between the devices that lead to wasteful re-transmissions, draining the battery life of low-power devices significantly. We present CharIoT, the first cross-technology distributed MIMO receiver system that exploits the potential of distributed MIMO to facilitate better co-existence and decoding of a large number of simultaneous low power uplink transmissions from unmodified low-power clients. CharIoT is a recovery-based system that intelligently collects radio samples from teams of light-weight IoT gateways and streams them to the cloud to effectively resolve collisions. At the cloud, CharIoT develops a suite of technology-specific software filters that decouple collisions across diverse technologies, facilitating seamless co-existence across low power radios. An implementation of CharIoT on inexpensive RTL-SDR gateways connected to Raspberry Pis decode collisions of four popular IoT technologies in the 868MHz ISM bands - LoRa, XBee, Z-Wave, and SIGFOX showing gains in throughput of up to 4x and battery life of up to 3.5 years.

To meet the ever increasing data rate demand in cellular network, the 3rd Generation Partnership Project (3GPP) introduced Licensed Assisted Access (LAA) with Listen Before Talk (LBT) mechanism to utilize the unlicensed spectrum to supplement the licensed LTE users. However, this leads to interference to already incumbent technologies in unlicensed spectrum such as WiFi. Furthermore, multi-carrier aggregation has been proposed to further increase the capacity in unlicensed spectrum. As WiFi primarily works on the principle of energy detection in the channel, in this paper, we first develop an analytical model to measure the performance of LAA and WiFi coexistence in unlicensed spectrum which supports multi-carrier aggregation which employs SINR based collision that allows the possibility of concurrent transmissions. Further, to reduce the interference to WiFi, we propose a coexistence method named Carrier Aggregation with Transmit Power Control (CATPC) which dynamically adjusts the LAA transmit power. CATPC provides a better opportunity for WiFi devices, better channel utilization, and energy efficient transmission by LAA. Through extensive simulations we demonstrate the effectiveness of CATPC when compared to other coexistence methods proposed in the literature.

Network Functions Virtualization (NFV) allows flexibility, scalability, agility, and easy manageability of networks by leveraging the features of virtualization and cloud computing technologies. However, softwarization of network functions imposes many challenges. Reliability and latency are major challenges in NFV-enabled 5G networks that can lead to customer dissatisfaction and revenue loss. In general, redundancy is used to improve the reliability of communication services. However, redundancy requires the same amount of additional resources and thus increases cost. In this article, we address the reliability-aware, delay guaranteed, and resource efficient Service Function Chain (SFC) placement problem in softwarized 5G networks. First, we propose a novel SFC subchaining method to enhance the reliability of an SFC without backups. If reliability requirement is not met after subchaining method, we add backups to VNFs to meet the reliability requirement. Then, we formulate the reliable SFC placement problem as an Integer Linear Programming (ILP) problem in order to solve it optimally. Owing to high computational complexity of the ILP problem for solving large input instances, we propose a modified stable matching algorithm to provide near-optimal solution in polynomial time. By extensive simulations we show that our proposed solutions consume lesser physical resources compared to state-of-the-art solutions for provisioning reliable communication services.