Latency Data Research Articles

Resource Allocation of Device‐To‐Device–Enabled Millimeter‐Wave Communication: A Deep Reinforcement Learning Approach

ABSTRACTDevice‐to‐device (D2D) communication is a promising development in 5G networks, offering potential benefits such as increased data rates, reduced costs and latency, and improved energy efficiency (EE). This study analyzes the operation of millimeter‐wave (mmWave) in cellular networks. A client's device can establish a connection to either a base station or another client, facilitating D2D communication based on a distance threshold and accounting for interference. The research employs a deep reinforcement learning (DRL)–based resource allocation (RA) scheme for D2D‐enabled mmWave communications underlaying cellular networks. It evaluates the effectiveness of several metrics: coverage probability, area spectral efficiency, and network EE. Among networks limited by noise, the proposed strategy demonstrates the highest coverage probability performance. The paper also suggests an optimization approach based on the firefly algorithm for RA, taking into account the stochastic nature of wireless channels. An asynchronous advantage actor–critic (A3C) DRL algorithm is modeled for this purpose. The performance of the proposed scheme is compared with two existing algorithms: soft actor–critic and proximal policy optimization. Overall, the numerical results indicate that our proposed firefly algorithm–optimized A3C method outperforms the other analytical methods.

Enhancing financial portfolio management with predictive analytics and scalable data modeling techniques

The financial industry faces an increasingly complex landscape for portfolio management, where data-driven insights are crucial for optimizing asset allocation and managing risk. This paper explores the integration of predictive analytics and scalable data modeling techniques in enhancing financial portfolio management. By leveraging machine learning algorithms, big data architectures, and real-time data processing, predictive analytics can forecast asset trends, detect market anomalies, and assess portfolio risk with high precision. We evaluate several predictive models, including time-series forecasting, neural networks, and ensemble methods, for their efficacy in financial prediction. The study also discusses the role of scalable data modeling frameworks, such as Apache Spark and cloud-based data lakes, in handling vast volumes of unstructured data across different markets and asset classes. Findings indicate that predictive analytics, when paired with robust data models, can deliver real-time, actionable insights, enhancing decision-making for fund managers and institutional investors. Furthermore, this paper highlights best practices for implementing scalable models in financial institutions, addressing challenges like data latency, model interpretability, and system scalability. By adopting these advanced analytics frameworks, portfolio managers can achieve improved risk-adjusted returns, better asset diversification, and enhanced adaptability to market volatility.

Hybrid Algorithm Combining Bellman-Ford, Dijkstra, and Machine Learning for Dynamic Network Routing

Abstract: Optimizing routing in dynamic networks has emerged as an increasingly important problem today with re- duced latency, guaranteed connectivity and efficient data trans- fer under a variety of changing conditions. Among the more well known shortest path algorithms those of Bellman-Ford and Dijkstra’s have very notable advantages and disadvantages themselves; while Bellman-Ford has provisions against negative weighted edges and pays a penalty in computation, Dijkstra’s algorithm cannot work with negative weighted edges but is highly efficient. It introduces a new hybrid routing algorithm that makes intelligent choices between the classical Bellman-Ford and Dijk- stra’s using machine learning techniques. This model, through historical analysis of data on traffic data and network metrics, predicts the level of congestion and identifies that algorithm which is most efficient in the selection of a call for routing. It switches over to Bellman-Ford if congestion expectation is high or if it contains negative weights, whereas under stable network conditions, it employs Dijkstra’s due to its efficiency. Preliminary experiments reveal that this adaptive strategy optimizes both computation overhead and route selection accuracy in large scale dynamic networking environments. Our results indicate improvements in routing efficiency in complex networks based on predictive analytics from established algorithms

The crucial role of 5G, 6G, and fiber in robotic telesurgery.

This paper explores the role of 5G-and future 6G networks-in advancing robotic telesurgery by minimizing latency and enhancing data reliability for real-time remote operations. With robotic telesurgery gaining prominence as a tool to democratize access to specialized surgical care, telecommunications infrastructure has become central to its feasibility and safety. Key elements include 5G's capacity for ultra-low latency and high data transfer rates, which support critical modalities such as kinesthetic, audiovisual, and tactile feedback in telesurgery. The paper outlines the differing latency demands of these modalities, noting that kinesthetic data are particularly sensitive, requiring ultra-low latency for effective surgeon feedback. In addition, the paper discusses the importance of network reliability and Quality-of-Service (QoS) agreements, alongside the potential for 6G networks to further reduce latency and integrate AI-driven predictive analytics. These advancements are positioned to not only broaden telesurgery's reach but also to enhance the precision and safety of procedures, setting the stage for a new paradigm in remote surgical care.

Data management solutions for real-time analytics in retail cloud environments

In today's rapidly evolving retail landscape, the importance of real-time data analytics has surged, driven by the need for businesses to make informed decisions and enhance customer experiences. This paper explores the critical role of effective data management solutions within cloud environments, focusing on retailers' unique challenges, such as data integration, latency, security, and storage costs. It presents practical solutions, including data partitioning, caching, AI-driven analytics, and the strategic use of data warehousing and data lakes. By adopting hybrid data management models, retailers can optimize their data processes and leverage real-time insights to drive operational efficiency and competitive advantage. This paper offers valuable recommendations for retail businesses and data architects to enhance their cloud-based data management strategies, ensuring they can navigate the complexities of real-time data effectively and sustainably. Keywords: Real-time analytics, Data management, Retail, Cloud environments, Data integration, AI-driven solutions.

Evaluating the efficacy of telesurgery with dual console SSI Mantra Surgical Robotic System: experiment on animal model and clinical trials

The field of robotic surgery has grown exponentially over the past few decades. Surgical robots offer numerous benefits that enhance surgical precision, improve patient outcomes, and expand the capabilities of surgeons. Telesurgery, also known as a remote surgery, is a branch of telemedicine, which offers to perform surgical procedures requiring expertise of a surgeon located at a distance from a patient by using robotic systems and telecommunications technology. In a previous reported case, an animal experiment and clinical trial telesurgery using a dual console were performed. However, the mean latency time and data packet loss were considerably high. As a result, the performance of the telesurgery got severely impacted. This paper evaluates the feasibility, safety, and efficacy of remote telesurgery to precisely carryout predetermined surgical procedures using dual console SSI Mantra Surgical Robotic System. The trials were registered prospectively with trial registration number CTRI-2024-06-068361.

RENEWED: A follow-up study of the opicinumab phase 2 RENEW study in participants with acute optic neuritis

BackgroundThe randomized, phase 2 RENEW trial (NCT01721161) evaluated efficacy/safety of opicinumab (anti-LINGO-1) versus placebo in patients with first-episode unilateral acute optic neuritis (AON). Although no significant differences in the latency recovery of visual evoked potential (VEP) were observed between opicinumab and placebo groups in the intention to treat (ITT) population, the prespecified per-protocol (PP) population showed better recovery with opicinumab than with placebo. RENEWED (NCT02657915) was a one-visit, follow-up study 2 years after the last RENEW study visit (Week 32) designed to assess the long-term electrophysiological and clinical outcomes for participants previously enrolled and having received study treatment in RENEW. MethodsIn the original study (RENEW), participants (aged 18–55 years) with a first unilateral AON episode were enrolled ≤28 days from first symptom onset and after treatment with methylprednisolone 1 g/day intravenously for 3–5 days; these participants were randomized to receive opicinumab 100 mg/kg or placebo intravenously once every 4 weeks from baseline to Week 20, assessed up to Week 32. Participants who received ≥1 dose of opicinumab 100 mg/kg or placebo in RENEW were eligible for the RENEWED follow-up study. Participants enrolled in RENEWED at 2 years (with an additional up to 12-month window) after the last RENEW study visit (Week 32) in both ITT and PP populations. The primary endpoint was change in full-field VEP (FF-VEP) latency of the affected eye at RENEWED study visit versus baseline of the fellow eye in RENEW, comparing between participants who received opicinumab and placebo in RENEW. Clinical progression and severity of multiple sclerosis (MS) were assessed. A substudy evaluated latency recovery using multifocal VEP (mfVEP) as an exploratory endpoint. ResultsOf 82 RENEW participants, 52 (63.4%; opicinumab n=28, placebo n=24) enrolled in and completed RENEWED. The adjusted mean (95% CI) difference in FF-VEP latency delay between opicinumab and placebo groups was −6.0 (−14.6, 2.6) msec (p=0.165) for the PP population and −4.5 (−12.6, 3.7) msec (p=0.274) for the ITT population at the RENEWED study visit. Nominally significant improvement on mfVEP latency in the opicinumab group versus placebo was observed in participants of the mfVEP substudy (p=0.009). In participants from the PP population without clinically definite MS (CDMS) at RENEW baseline,12 (55%) in the opicinumab group and 12 (67%) in the placebo group developed CDMS from enrollment in the RENEW study up to RENEWED Day 1; the estimated proportion of participants with CDMS at 2 years after the last study visit assessment in RENEW was lower when treated with opicinumab (0.50) than when treated with placebo (0.61) (hazard ratio p-value=0.23). No benefit on visual acuity or other neurological functions was observed in the opicinumab group vs placebo in RENEWED. ConclusionThe numerically increased VEP latency recovery with opicinumab treatment in RENEWED was consistent with those observed in the parent study RENEW. However, the VEP latency and clinical data in RENEWED should be interpreted with caution, given the nature of the follow-up study, the small sample size and the limitation in study design.

Scalable SOA-based Banyan optical space switch on InP membrane on silicon (IMOS).

Optical switches (OS) are vital for meeting modern data centers' capacity and latency needs, overcoming the bandwidth and speed limitations that electronic switches encounter. To this end, this work proposes a semiconductor optical amplifier (SOA)-based OS on the IMOS, a unique platform with a combination of high refractive index contrast for compactness and monolithic active-passive integration for active functionalities, thereby enabling the creation of compact SOA-based OSes. An 8 × 8 Banyan SOA-based OS is integrated on a 4 × 4 mm2 area on the standard 4 × 6 mm2 IMOS cell. This marks the first application of the IMOS technology platform for SOA-based OSes and sets the stage for compact and large-scale monolithic photonic integrated OS circuits. The basic 2 × 2 switch module delivers a high optical signal-to-noise ratio (OSNR) and an extinction ratio (ER) above 45 dB. Employing NRZ-OOK routing on the 2 × 2 basic OS module, a 15 dB input power dynamic range (IPDR) is achieved within a 1 dB power penalty at 12.5 Gb/s. Data routing at higher data rates of 25 Gb/s and 40 Gb/s incurs power penalties of 0.8 and 1.4 dB, respectively. Furthermore, data routing for a 3-stage 8 × 8 switch operating at 25 Gb/s results in a power penalty of 1.2 dB. Compared to the 1-stage 2 × 2 switch, only a 0.4 dB additional penalty is observed despite incorporating 2 additional SOAs in the cascade. This indicates that the switch can scale to higher radix to accommodate multi-stage switches with numerous ports, large bandwidth, and fast speed, which is essential in modern large-scale data centers.

Regressive Machine Learning for Real-Time Monitoring of Bed-Based Patients

This study introduces an ensemble model designed for real-time monitoring of bedridden patients. The model was developed using a unique dataset, specifically acquired for this study, that captures six typical movements. The dataset was balanced using the Synthetic Minority Over-sampling Technique, resulting in a diverse distribution of movement types. Three models were evaluated: a Decision Tree Regressor, a Gradient Boosting Regressor, and a Bagging Regressor. The Decision Tree Regressor achieved an accuracy of 0.892 and an R2 score of 1.0 on the training dataset, and 0.939 on the test dataset. The Boosting Regressor achieved an accuracy of 0.908 and an R2 score of 0.99 on the training dataset, and 0.943 on the test dataset. The Bagging Regressor was selected due to its superior performance and trade-offs such as computational cost and scalability. It achieved an accuracy of 0.950, an R2 score of 0.996 for the training data, and an R2 score of 0.959 for the test data. This study also employs K-Fold cross-validation and learning curves to validate the robustness of the Bagging Regressor model. The proposed system addresses practical implementation challenges in real-time monitoring, such as data latency and false positives/negatives, and is designed for seamless integration with hospital IT infrastructure. This research demonstrates the potential of machine learning to enhance patient safety in healthcare settings.

A Performance Benchmark for the PostgreSQL and MySQL Databases

This study highlights the necessity for efficient database management in continuous authentication systems, which rely on large-scale behavioral biometric data such as keystroke patterns. A benchmarking framework was developed to evaluate the PostgreSQL and MySQL databases, minimizing repetitive coding through configurable functions and variables. The methodology involved experiments assessing select and insert queries under primary and complex conditions, simulating real-world scenarios. Our quantified results show PostgreSQL’s superior performance in select operations. In primary tests, PostgreSQL’s execution time for 1 million records ranged from 0.6 ms to 0.8 ms, while MySQL’s ranged from 9 ms to 12 ms, indicating that PostgreSQL is about 13 times faster. For select queries with a where clause, PostgreSQL required 0.09 ms to 0.13 ms compared to MySQL’s 0.9 ms to 1 ms, making it roughly 9 times more efficient. Insert operations were similar, with PostgreSQL at 0.0007 ms to 0.0014 ms and MySQL at 0.0010 ms to 0.0030 ms. In complex experiments with simultaneous operations, PostgreSQL maintained stable performance (0.7 ms to 0.9 ms for select queries during inserts), while MySQL’s performance degraded significantly (7 ms to 13 ms). These findings underscore PostgreSQL’s suitability for environments requiring low data latency and robust concurrent processing capabilities, making it ideal for continuous authentication systems.

Blockchain-machine learning fusion for enhanced malicious node detection in wireless sensor networks

In wireless sensor networks (WSNs), the presence of malicious nodes (MNs) poses significant challenges to data integrity, network stability, and system reliability. These issues are intensified by energy resource constraints and limitations within centralized authentication systems, necessitating an energy-efficient solution to ensure real-time responsiveness. Although artificial intelligence-driven approaches enhance detection capabilities, they overcome challenges related to data volume, coordination overhead, and latency in centralized control. This study introduces blockchain-machine learning (BC-ML), a novel hybrid model that seamlessly integrates blockchain and machine learning (ML) techniques to effectively identify MNs in WSNs. The model establishes an energy-efficient blockchain among cluster heads (CHs) for robust node authentication, incorporating a Schnorr-like zero-knowledge-proof technique to validate node data during communication initiation. Utilizing a hybrid lightweight approach with both symmetric and asymmetric ciphers enhances the security of node data transmission. A new proof-of-authority method is introduced, which leverages node digital certificates instead of conventional data transactions. This consensus mechanism reduces the processing overhead associated with larger data sizes in traditional proof-of-work methods, thereby improving both energy efficiency and scalability. To address dataset imbalances, the model employs a hybrid unsupervised ML technique, combining adaptive synthetic sampling with a convolutional neural network for efficient analysis of nodes and network features. The ML model, hosted on a robust data server, ensures ongoing oversight by updating CHs with security levels for detected MNs, thereby reducing storage and mitigating coordination challenges. Comprehensive analyses validate the effectiveness of the BC-ML model for detecting MNs, optimizing resource utilization, minimizing delays, and prolonging node and network lifetimes. Security analysis further confirms the ability of the model to mitigate diverse attacks and meet the stringent WSN security requirement.

Implementing Linkage Quality Assessments in Large-Scale Privacy Preserving Infrastructure: Did we kill them a second time and why are there zombies?

The National Clinical Cohort Collaborative (N3C) uses privacy-preserving record linkage (PPRL) to formulate longitudinal patient histories and for de-duplication across disparate data including electronic health records, administrative claims, and mortality data. The N3C employs a linkage honest broker implementation model which separates the entity holding de-identified PPRL tokens, from the entity holding the substantive data payloads. As a result of this data governance structure, linkage quality assessments are necessarily distributed into 2 distinct processes: 1) Linkage Honest Broker receipt and processing of linkages, 2) Data Aggregation entity assembling data based on received linkage crosswalks. Data is released on a weekly basis. This presentation will focus on Step 2 for the Mortality data linkages where the downstream data linkage pipelines must assess linkage quality across several dimensions of quality, focused on plausibility. There are 4 types of mortality data spanning the electronic health record, government Social Security Administration death master files, third-party obituary data, and administrative claims. Each of these sources has different levels of authoritativeness and data latency, compounding the challenges of reconciling fact of death including accurate death reporting dates. To address these challenges, we implemented an automated linkage quality pipeline which generates reports that assess linkage plausibility across the various data types, classifying concordance and discordance between data sources including age, dates of death, surfacing plausible and non-plausible activities post-death. Sharing our real-world implementation will aid conference attendees in understanding how to plan for such activities when considering specific implementation models for privacy-preserving linkages.

Access control for trusted data sharing

In the envisioned 6G landscape, data sharing is expected to become increasingly prevalent, giving rise to digital marketplaces that foster cooperation among organizations for collecting, sharing, and processing data for analysis. These marketplaces serve as connectors between data producers and consumers, empowering multi-tenancy scenarios for seamless and secure data sharing both within and outside organizations. Given that 6G networks promise ultra-low latency, enhanced connectivity, and massive data throughput, the need for robust data access control mechanisms becomes imperative. These mechanisms ensure security and trust among entities, particularly in multi-tenant environments where multiple organizations share infrastructure and data resources. In this paper, we have designed and implemented a novel access control mechanism tailored for a distributed data streaming system developed by Nokia Bell Labs. Our approach leverages fine-grained policies, dynamic enforcement, and transparency mechanisms to enhance trust between data owners and consumers. By facilitating secure multi-tenancy data sharing, our solution contributes to the seamless exchange of data across diverse entities within the next-generation communication ecosystem. We demonstrate that our proposed access control mechanism incurs minimal overhead while ensuring data confidentiality and integrity. The introduction of such advancements in data sharing markets strengthens the overall ecosystem by providing heightened transparency and enhanced control over data, promoting collaboration and innovation in the 6G era.

Blockchain-based security-minded information-sharing in precast construction supply chain management with scalability, efficiency and privacy improvements

Blockchain and Interplanetary File System (IPFS) integration holds great promise for enhancing transparency and traceability in precast construction supply chain management (PCSCM). However, such integration faces challenges regarding unauthorized access to confidential data, which can lead to significant consequences, such as financial losses and legal issues. Regarding this gap, this paper proposes a hybrid privacy-preserving access control mechanism for securing blockchain-IPFS information-sharing in PCSCM with confidentiality considerations. Multiple privacy-preserving techniques (e.g., information-sharing channels, symmetric and asymmetric encryption, and lightweight proxy re-encryption) are leveraged. A prototype system demonstrates its feasibility and effectiveness, achieving demonstrably low network latency (millisecond level), efficient encryption (millisecond level), and robust data security within both blockchain and IPFS. This research contributes to a deeper understanding of blockchain-IPFS integration and provides a valuable reference point for future research and practical adoption.

Evaluating Dimensionality Reduction Methods for the Detection of Industrial IoT Attacks in Edge Computing

Edge computing is essential for 6G mobile networks for improving reliability, reducing data rates and latency, and enhancing mobile connectivity. Edge computing is also meant to meet the increasing demands of the Internet of Things (IoT)/ Internet of Everything (IoE). In these approaches, IIoT systems necessitate precision, reliability, and scalability, while vulnerabilities in IIoT systems may lead to financial losses and safety hazards. To tackle this, Edge AI/ML-based IDSs provide adaptability and robustness for IIoT security challenges. These solutions improve security levels, threat detection rates, and response times. However, main issues such as limited resources and the accuracy of attack detection rate are challenged nowadays. In this paper, we present contributions in proposing an intrusion detection system (IDS) for edge devices deploying a lightweight deep neural network to detect IIoT attacks. We present performance analysis of the typical dimensionality reduction methods and balancing data features of the Edge-IIoT dataset to get higher performance metrics than other state-of-the-art studies.

An Improved Scoring Algorithm for Indirect Evaluation Measurement With the Evaluative Priming Task

People tend to categorize the valence of a target stimulus more quickly and accurately if the target appears after a prime stimulus of the same (than of the opposite) valence. The evaluative priming task (EPT) utilizes this priming effect for indirect evaluation measurement but suffers from low reliability. The present research compared the reliability and validity of 2,160 EPT scoring algorithms across 12 datasets. In contrast to current norms to delete trials with error responses and to rely solely on differences in response latency between different task conditions, superior performance was found when incorporating latency and accuracy data, and when basing the score on differences in mean ranking of each condition’s trials by performance (a scale-invariant non-parametric dominance score named G score). We recommend seven new scoring algorithms that, in comparison to current scoring norms, increase internal consistency by a mean of 41% and correlations with other measures by 17%.

Real-time data warehousing in the big data environment: A comprehensive review of implementation in the internet industry

A real-time data warehouse is a crucial tool for information management and analysis, enabling the capture, processing, and analysis of vast amounts of data from diverse sources in real-time. It offers enterprises enhanced decision support through its efficient processing capabilities and timely data feedback. This paper reviews the technical characteristics and application scenarios of real-time data warehouses, with a particular focus on the Internet sector. It explores the evolution from traditional data warehouses to modern data lake and lakehouse architectures, emphasizing the advancements in data processing capabilities, including the separation of storage and compute functions. Real-time data warehouses, which enable immediate data processing and feedback, are essential for enterprises requiring up-to-the-minute insights. The study compares the Lambda and Kappa architectures, detailing their strengths and weaknesses in terms of data throughput, latency, and scalability. Innovations such as Apache Hudi and lakehouse architectures offer new opportunities for performance optimization and functional expansion. The emergence of hybrid architectures like HTAP (Hybrid Transactional/Analytical Processing) and HSAP (Hybrid Serving/Analytical Processing) represents a significant advancement in integrating transactional and analytical processing. Future research should focus on the impact of artificial intelligence and machine learning on real-time data warehouses to enhance their analytical and predictive capabilities, reduce complexity, and lower operational costs.

From Near-Sensor to In-Sensor: A State-of-the-Art Review of Embedded AI Vision Systems.

In modern cyber-physical systems, the integration of AI into vision pipelines is now a standard practice for applications ranging from autonomous vehicles to mobile devices. Traditional AI integration often relies on cloud-based processing, which faces challenges such as data access bottlenecks, increased latency, and high power consumption. This article reviews embedded AI vision systems, examining the diverse landscape of near-sensor and in-sensor processing architectures that incorporate convolutional neural networks. We begin with a comprehensive analysis of the critical characteristics and metrics that define the performance of AI-integrated vision systems. These include sensor resolution, frame rate, data bandwidth, computational throughput, latency, power efficiency, and overall system scalability. Understanding these metrics provides a foundation for evaluating how different embedded processing architectures impact the entire vision pipeline, from image capture to AI inference. Our analysis delves into near-sensor systems that leverage dedicated hardware accelerators and commercially available components to efficiently process data close to their source, minimizing data transfer overhead and latency. These systems offer a balance between flexibility and performance, allowing for real-time processing in constrained environments. In addition, we explore in-sensor processing solutions that integrate computational capabilities directly into the sensor. This approach addresses the rigorous demand constraints of embedded applications by significantly reducing data movement and power consumption while also enabling in-sensor feature extraction, pre-processing, and CNN inference. By comparing these approaches, we identify trade-offs related to flexibility, power consumption, and computational performance. Ultimately, this article provides insights into the evolving landscape of embedded AI vision systems and suggests new research directions for the development of next-generation machine vision systems.

Review on LoRa Communication Technology, Its Issues, Challenges and Applications in Healthcare System

The Internet of Things (IoT) has transformed various industries by enabling interconnected devices to collect, share, and analyze data in real-time. A crucial component of this transformation is Long Range (LoRa) technology, designed for low-power, wide-area networks (LPWAN). LoRa enables long-distance communication with minimal energy consumption, making it ideal for IoT applications in sectors like healthcare, agriculture, and smart cities. Specifically, in healthcare, LoRa facilitates remote monitoring through the Internet of Medical Things (IoMT), where patient data, such as vital signs, can be transmitted efficiently over vast distances. This paper discusses the significance of LoRa in medical applications, its advantages, challenges, and solutions, and reviews existing literature on its implementation. Key challenges include low data rates, packet loss, and latency issues, while solutions such as adaptive data rate mechanisms and multi-hop networks offer potential improvements for medical IoT systems. Also based on the given comparison among communication technologies, the consumer can make his decision on chosen the right technology for his application.

Comparison of Reinforcement Learning Algorithms for Edge Computing Applications Deployed by Serverless Technologies

Edge computing is one of the technological areas currently considered among the most promising for the implementation of many types of applications. In particular, IoT-type applications can benefit from reduced latency and better data protection. However, the price typically to be paid in order to benefit from the offered opportunities includes the need to use a reduced amount of resources compared to the traditional cloud environment. Indeed, it may happen that only one computing node can be used. In these situations, it is essential to introduce computing and memory resource management techniques that allow resources to be optimized while still guaranteeing acceptable performance, in terms of latency and probability of rejection. For this reason, the use of serverless technologies, managed by reinforcement learning algorithms, is an active area of research. In this paper, we explore and compare the performance of some machine learning algorithms for managing horizontal function autoscaling in a serverless edge computing system. In particular, we make use of open serverless technologies, deployed in a Kubernetes cluster, to experimentally fine-tune the performance of the algorithms. The results obtained allow both the understanding of some basic mechanisms typical of edge computing systems and related technologies that determine system performance and the guiding of configuration choices for systems in operation.