Orchestrating Distributed Edge Resources with Cloud Native Frameworks

Edge computing represents a transformative paradigm shift for real-time media applications, fundamentally altering how processing resources are distributed across network infrastructures. This article examines the evolution from centralized cloud architectures to distributed edge computing models, addressing critical challenges in latency reduction, bandwidth optimization, and scalability for media-intensive applications. Through distributed processing topologies, edge-cloud integration frameworks, and data flow optimization techniques, we present quantitative performance improvements achieved through edge deployment. The article explores latency reduction methodologies, intelligent data distribution strategies, and scalability solutions that collectively enhance media delivery across diverse application domains. Case studies in live event broadcasting, video conferencing, smart surveillance, and telehealth demonstrate the practical benefits of edge-based media processing. Looking forward, we examine emerging trends including integration with next-generation networks, AI-enhanced media optimization, standardization efforts, and specialized hardware accelerators that will shape the future landscape of edge computing for real-time media applications.

<p>Internet of Things (IoT) requires cloud infrastructures for data analysis (e.g., temperature monitoring, energy consumption measurement, etc.). Traditionally, cloud services have been implemented in large datacenters in the core network. Core cloud offers high-computational capacity with moderate response time, meeting the requirements of centralized services with low-delay demands. However, collecting information and bringing it into one core cloud infrastructure is not a long-term scalable solution, particularly as the volume of IoT devices and data is forecasted to explode. A scalable and efficient solution, both at the network and cloud level, is to distribute the IoT analytics between the core cloud and the edge of the network (e.g., first analytics on the edge cloud and the big data analytics on the core cloud). For an efficient distribution of IoT analytics and use of network resources, it requires to integrate the control of the transport networks (packet and optical) with the distributed edge and cloud resources in order to deploy dynamic and efficient IoT services. This paper presents and experimentally validates the first IoT-aware multilayer (packet/optical) transport software defined networking and edge/cloud orchestration architecture that deploys an IoT-traffic control and congestion avoidance mechanism for dynamic distribution of IoT processing to the edge of the network (i.e., edge computing) based on the actual network resource state.</p>

Long range (LoRa) is a wireless communication standard specifically targeted for resource-constrained Internet of Things (IoT) devices. LoRa is a promising solution for smart city applications as it can provide long-range connectivity with a low energy consumption. The number of LoRa-based networks is growing due to its operation in the unlicensed radio bands and the ease of network deployments. However, the scalability of such networks suffers as the number of deployed devices increases. In particular, the network performance drops due to increased contention and interference in the unlicensed LoRa radio bands. This results in an increased number of dropped messages and, therefore, unreliable network communications. Nevertheless, network performance can be improved by appropriately configuring the radio parameters of each node. To this end, in this article we formulate integer linear programming models to configure LoRa nodes with the optimal parameters that allow all devices to reliably send data with a low energy consumption. We evaluate the performance of our solutions through extensive network simulations considering different types of realistic deployments. We find that our solution consistently achieves a higher delivery ratio (up to 8% higher) than the state of the art with minimal energy consumption. Moreover, the higher delivery ratio is achieved by a large percentage of nodes in each network, thereby resulting in a fair allocation of radio resources. Finally, the optimal network configurations are obtained within a short time, usually much faster than the state of the art. Thus, our solution can be readily used by network operators to determine optimal configurations for their IoT deployments, resulting in improved network reliability.

Named Data Networking (NDN) revolutionizes content-centric communication through intrinsic in-network caching, yet optimal cache management remains challenging under dynamic content popularity and resource constraints. This paper introduces GT-DRLCache, a novel framework that synergistically integrates non-cooperative game theory with multi-agent deep reinforcement learning (MADRL) for intelligent caching decisions. We formulate the distributed caching problem as a stochastic congestion game with provable Nash equilibrium existence. Each router employs a hybrid Double-DQN/Proximal Policy Optimization (PPO) agent that learns equilibrium-aware caching strategies while adapting to temporal traffic patterns. Extensive simulations using ndnSIM 2.8 on GEANT, Abilene, and synthetic Barabási-Albert topologies demonstrate consistent performance improvements: 32–56% higher hit ratios, 28–44% latency reduction, and up to 50% upstream traffic savings compared to LRU, LFU, MPC, ProbCache, and DRL-only baselines. This work provides comprehensive theoretical foundations, complete algorithmic implementation, and thorough experimental validation of the proposed unified caching framework.

In the context of Internet of Things (IoT), optimizing quality of service (QoS) parameters is a critical challenge due to its heterogeneous and resource-constrained nature. This paper proposes a novel quantum-inspired multi-objective optimization algorithm for IoT service management. Traditional multi-objective optimization algorithms often face limitations such as slow convergence and susceptibility to local optima, reducing their effectiveness in complex IoT environments. To address these issues, we introduce a quantum-inspired hybrid algorithm that combines the strengths of Multi-Objective Grey Wolf Optimization Algorithm (MOGWOA) and Multi-Objective Whale Optimization Algorithm (MOWOA), enhanced with quantum principles. This novel integration overcomes the limitations of traditional algorithms by improving convergence speed and avoiding local optima. The hybrid algorithm enhances QoS in IoT applications by achieving superior optimization in terms of energy efficiency, latency reduction, convergence, and coverage cost. The incorporation of quantum-inspired mechanisms, such as quantum position and behavior, strengthens the exploration and exploitation capabilities of the algorithm, enabling faster and more accurate optimization. Extensive simulations and testing demonstrate the proposed method’s superior performance compared to existing algorithms, validating its effectiveness in addressing key IoT challenges.

Self-aware collaborative edge inference with embedded devices for IIoT

Joint route guidance and demand management using generalized MFDs

We introduce a smart multi-objective optimization framework towards the performance enhancement of the 4G LTE network using an integrated Artificial Intelligence (AI) and Whale Optimization Algorithm (WOA) setup. This approach aims to integrate tested regression-based KPI prediction along with decision-variable-based meta-heuristic optimization in a reproducible experimental environment within MATLAB R2023b. A weighted multi-objective formulation for optimizing the performance of 4 key indicators, i.e., throughput, latency, packet loss ratio, and energy efficiency, is jointly executed. By comparison against existing baseline LTE configuration, we observe a 25.3% throughput improvement, 28.9% reduction in latency, 38.7% reduction in packet loss ratio, and 23.4% improvement in energy efficiency. We validate robustness with weight sensitivity analysis (±10% and ±20%), population scalability testing (N = 20, 30, 50), decay-strategy comparison (linear vs. exponential), and Pareto-front approximation (2,960 non-dominated solutions from 3,000 samples). The above results corroborate that, after approximately 70 iterations, convergence is stable and performance is consistent over the span of 30 independent runs. Thus, the proposed AI–WOA construction offers an organized and replicable process to fine-tune LTE performance optimization, a basis for downstream implementations towards a more distributed and next-generation wireless environment.

Mobile wireless sensor control systems demand low-latency and energy-efficient data transmission to ensure timely decision-making in dynamic environments such as industrial automation, healthcare monitoring, and intelligent transportation. The conventional communication protocols, such as ZigBee offer energy efficiency but is not always effective in supporting tight latency constraints in high mobility and traffic congestion. The work introduces the architecture and analysis of a low-latency data transmission system with an optimized MQTT-SN protocol, augmented with the packet aggregation and priority-based scheduling, dynamic channel allocation and cross-layer optimization. The proposed system is verified by extensive MATLAB simulations and testbed experiments on ESP32. Findings show that the reduction in latency is up to 60% over that of ZigBee with an average delay of 18.6 ms at 10 nodes and 69.1 ms at 100 nodes. Throughput is increased by 10-15, and MQTT-SN with optimization attains 430.7 kbps at 500 kbps load or less. A packet delivery ratio (PDR) of over 97% is ensured at low mobility and 86.9% even at <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$5 ~\mathrm{m} / \mathrm{s}$</tex> mobility speeds. Consumption of energy per packet is cut by 25 to 30% and jitter is less than 15 ms, guaranteeing predictable performance of control. The general accuracy of the system expressed as the percentage of successful and timely packet transmissions had an average of 96.4% which was higher than the 89.5% observed in the ZigBee system. The solution proposed has better control of latency, reliability and efficiency, which has made it very appropriate to use in applications of mobile sensors that require latency control. Such results indicate its capability to be applied in actual smart environments.

Ensuring ultra-low latency and high reliability in 6G network slices remains a significant challenge, as current NFV orchestration approaches are largely reactive and not designed to anticipate performance degradation. The advent of 6G networks brings forth stringent requirements for ultra-reliable low-latency communication (URLLC), necessitating advanced orchestration mechanisms that go beyond reactive policies in traditional NFV environments. In this paper, we propose a latency-aware, AI-driven NFV slice orchestration framework aligned with ETSI MANO architecture to address the needs of time-sensitive 6G applications. Our framework integrates a predictive AI engine into the NFV Orchestrator (NFVO) to forecast latency violations based on real-time telemetry and historical trends. It enables dynamic scaling, intelligent VNF migration, and infrastructure-level isolation to maintain stringent end-to-end (E2E) latency targets. Experimental results indicate up to 30% reduction in average latency, a 42% improvement in SLA compliance, and 25% lower migration overhead compared to traditional reactive orchestration. The framework provides a scalable and intelligent orchestration solution adaptable to future 6G deployments.

Due to the capacity of processing signal with low energy consumption, compressive sensing (CS) has been widely used in wearable health monitoring system for arrhythmia classification of electrocardiogram (ECG) signals. However, most existing works focus on compressive sensing reconstruction, in other words, the ECG signals must be reconstructed before use. Hence, these methods have high computational complexity. In this paper, the authors propose a cardiac arrhythmia classification scheme that performs classification task directly in the compressed domain, skipping the reconstruction stage. The proposed scheme first employs the Pan–Tompkins algorithm to preprocess the ECG signals, including denoising and QRS detection, and then compresses the ECG signals by CS to obtain the compressive measurements. The features are extracted directly from these measurements based on principal component analysis (PCA), and are used to classify the ECG signals into different types by the proposed semi-supervised learning algorithm based on support vector machine (SVM). Extensive simulations have been performed to validate the effectiveness of the proposed scheme. Experimental results have shown that the proposed scheme achieves an average accuracy of [Formula: see text] at a sensing rate of 0.7, compared to an accuracy of [Formula: see text] for noncompressive ECG data.

Energy efficiency is of utmost importance for wireless sensor networks deployed without any possibility of battery replenishments. Thus, design of energy efficient algorithms and protocols must consider resource constraints while maintaining the desired level of QoS. In this paper, we present EE-MAC, an Energy Efficient medium access control (MAC) protocol for distributed wireless sensor networks. EE-MAC achieves a low-duty-cycle and hence low energy consumption through optimized sleep intervals while transitioning between sleep and active states. We consider a weighted linear combination of delay and energy saving as the performance metrics and through extensive simulations, we observe reduced energy consumption at the cost of increased delay. EE-MAC also improves the delay performance for fixed number of nodes compared to S-MAC.

Microservice architectures (MSAs) have revolutionized software development by offering flexibility, scalability, and resilience through the decomposition of applications into loosely coupled services. However, resource management and performance optimization in MSAs remain challenging due to dynamic workloads and complex interdependencies. Traditional approaches, such as static provisioning and rule-based scaling, struggle to handle these challenges efficiently, often leading to over-provisioning or under-provisioning of resources. In this paper, we propose an AI-driven optimization framework that integrates reinforcement learning (RL), predictive analytics (PA), and evolutionary algorithms (EA) to dynamically manage resources in microservices environments. The proposed framework anticipates workload changes, optimizes resource allocation in real-time, and continuously adapts to evolving system conditions. Our empirical evaluation, conducted on a Kubernetes-based microservice platform, demonstrates significant improvements in performance and resource efficiency compared to conventional methods like Kubernetes' Horizontal Pod Autoscaler (HPA). The AI-driven system achieves up to a 27.3% reduction in latency during traffic surges and improves throughput by 25.7%, while also reducing CPU and memory usage by up to 25.7% and 22.7%, respectively. These results suggest that AI-driven optimization offers a scalable and efficient solution for managing microservices in highly dynamic environments.

Although collecting enormous volumes of heterogeneous data from many sensors and guaranteeing real-time decision-making are problems, precision agriculture (PA) has emerged as a promising approach to increase agricultural efficiency. The efficacy of current centralized solutions is limited in large-scale agricultural settings due to resource limitations and data saturation. In order to solve these problems, this paper suggests a decentralized computing environment for precision agriculture (DECPA), which divides resource management and data processing among several layers (end, edge, and cloud). DECPA optimizes task execution and resource allocation in the field by utilizing ensemble machine learning models (deep neural network (DNN), long short-term memory (LSTM), autoencoder (AE), and support vector machine (SVM)) and a multi-tier architecture. The findings demonstrate that DECPA combined with DNN performs better than alternative models, achieving a 20% decrease in energy usage, an 18% speedup in response time, a 5% improvement in accuracy, and a 51% reduction in latency. This illustrates the system’s capacity to manage massive amounts of data effectively while preserving peak performance. To sum up, DECPA uses decentralized resources and cutting-edge machine learning models to provide a scalable and affordable precision agriculture solution. To improve the system’s flexibility and real-time responsiveness, future research will investigate additional optimization and use in various agricultural contexts.

New water resources management methodologies are needed to address increasing demands and future uncertainty for urban water resources. Adaptive water demand management strategies provide an approach to improve the efficiency of water system operations and meet water demands by adapting flexibility to increasing stresses, such as droughts. This study simulates adaptive water demand management through the development of a complex adaptive system modeling framework, which couples cellular automata modeling, agent-based modeling, and hydrologic modeling to simulate land-use change, consumer behaviors, management decisions, the rainfall-runoff process, and reservoir storage. The model is applied to simulate the effect of demand management strategies on reductions in municipal water demands and on the sustained storage in a surface water supply reservoir. Historic and projected climate change hydroclimatic time series are used to assess the effectiveness of domestic water restrictions, including outdoor watering restrictions, a rainwater harvesting rebate program, and a high-density land-use change policy. Strategies are adaptively implemented based on the amount of water storage available. The framework is applied to evaluate strategies for the Arlington, Texas, metropolitan area that historically suffers from severe droughts. The framework provides an approach to evaluate a combination of multiple strategies for effectively managing the increasing stresses caused by urbanization, population growth, and climate change. Results demonstrate that adaptive demand management strategies that respond to water shortages result in long-term per capita demand reductions. For climate projections that forecast severe water shortages, development density strategies are more effective than rainwater harvesting strategies, and a combination of strategies can reduce the need for interbasin transfers and maintain reservoir volumes.