Federated Learning (FL) has emerged as a decentralised machine learning paradigm for distributed systems, particularly in edge and IoT environments. However, achieving fault tolerance and self-recovery in such scenarios is challenging due to the centralised model aggregation, which poses a single point of failure. This article focuses on the self-recovery of the aggregator, specifically the controlled re-assignment of the aggregator role to the most suitable node. Our proposed solution leverages the RAFT consensus algorithm to facilitate consistent state replication and leader election within the FL system. This is complemented by controlled aggregator re-assignment, which considers various contextual properties to select the optimal node, enhancing the system’s robustness, especially in dynamic and unreliable cyber-physical environments. We implement a proof of concept using the Flower FL framework and conduct experiments to evaluate aggregator recovery time and the traffic overhead associated with state replication. While the traffic overhead scales with the number of FL nodes, our results demonstrate a resilient, self-recovering system capable of handling node failures while maintaining model consistency.

The growing adoption of digital technologies in agriculture has led to a proliferation of heterogeneous data from sources such as drones, robotic platforms, and IoT sensors.However, the lack of interoperability across these data streams poses major challenges for integration into decision support systems.This paper presents an approach to harmonising such data using NGSI-LD and Smart Data Models, developed within the Norwegian research project SMARAGD.We demonstrate how domain-specific semantic models and linked data principles can be applied to standardise and enrich geospatial and temporal metadata across three key agritech domains: aerial imagery, robotic sensing, and environmental monitoring.The resulting information assets are integrated into a shared, FIWAREcompatible data space, enabling cross-platform visualisation, querying, and reuse.This work contributes to the development of an open, standards-based digital infrastructure for interoperable, data-driven agriculture in Norway and beyond.

The computing continuum aims to break the isolation of edge and cloud computing, creating a smooth and heterogeneous infrastructure surface for deploying applications. However, the continuum is practically fragmented, with infrastructure managed in isolation by various parties in a vertical dimension, i.e., edge, fog, and cloud layers, and horizontally, i.e., computing, network, and storage domains. This scenario negatively impacts the fulfillment of the application objectives. We propose inCoord, an intent-aware solution that enables the creation of a unified computing continuum by coordinating its instances to fulfill application objectives. Each instance represents a cluster of (compute, storage, or network) nodes with their own manager component. Traditionally, systems take low-level actions on these instances. In contrast, inCoord learns their emerging behaviors and adapts their managers’ objectives to fulfill the application’s intents. Here, through a Reinforcement-Learning-based Proof of Concept, we show the potential of this system to understand emerging behavior and manage multi-domain, independently managed instances.

ERATOSTHENES is an EC, co-funded, research project strongly considering modern security challenges in the domain of Internet of Things in mind of their huge penetration into our day to day lives. There are a series of recent challenges that recently have been converted into obstacles or risk points that could block the secure operation of IoT networks in all day to day activities, from home to office, to leisure and security. These include examples such as the highly increased number of connected devices (at all network levels) that are on top forming inhomogeneous networks and systems of systems. Different vendor characteristics further increase the attack surface that is expected to further rise in the upcoming years. Such, highly critical, characteristics, dramatically increase the needs for confidentiality access control, user and things’ privacy, devices’ trustworthiness and compliance that require lifecycle considerations. The ERATOSTHENES project orchestrates a novel distributed, automated, auditable, yet privacy-respectful, Trust and Identity Management Framework and Reference Architecture with the ultimate scope to dynamically and holistically manage IoT devices in a lifecycle approach, strengthening trust, identities, and resilience in the entire IoT ecosystem while supporting the enforcement of the NIS directive, GDPR and Cybersecurity Act. This publication describes the ERATOSTHENES technical concept and reference architecture as well as design considerations, architecture characteristics, connectivity and interoperability.

With increasing interdependencies and evolving threats, maintaining operational continuity in critical systems has become a significant challenge. This paper presents the DYNABIC (Dynamic business continuity of critical infrastructures on top of adaptive multi-level cybersecurity) approach as a comprehensive framework to enhance the resilience of critical infrastructures. The DYNABIC approach provides the resilience enhancement through dynamic adaptation, automated response, collaboration, risk assessment, and continuous improvement. By fostering a proactive and collaborative approach to resilience, the DYNABIC framework empowers critical infrastructure sectors to effectively mitigate disruptions and recover from incidents. The paper explores the key components and architecture of the DYNABIC approach and highlights its potential to strengthen the resilience of critical infrastructures using the concept of Digital Twins in the face of evolving threats and complex operating environments involving cascading effects.

Current security orchestration and response (SOAR) approaches have primarily focused on specific layers of systems, such as Intrusion Detection Systems, the network layer, or the application layer. We aim to find the gaps in the existing SOAR approaches for IoT/CPS-based systems, especially critical infrastructures, and propose some directions to fill in these gaps. This paper presents a literature survey and future research directions for advancing SOAR towards increased automation and more holistic operation, especially for the cyber-physical security of critical infrastructures. We have found 14 primary SOAR studies and discussed the gaps in general. There is a significant gap when it comes to a comprehensive and systematic approach to SOAR for multi-layered systems using IoT/CPS and considering the computing continuum perspective. To address the gap, we present our on-going work on a framework of multi-layer SOAR decision-making methods and orchestration tools that leverage Reinforcement Learning (RL)-based adaptation intelligence, virtual reality, avatar-human interaction and advanced Cyber Threat Intelligence (CTI) tools.

International audience

One of the most prominent implementations of the serverless programming model is Function-as-a-Service (FaaS). Using FaaS, application developers provide source code of serverless functions, typically describing only parts of a larger application, and define triggers for executing these functions on infrastructure components managed by the FaaS provider. There are still challenges that hinder the wider adoption of the FaaS model across the whole Cloud-Edge-IoT continuum. These include the high heterogeneity of the Edge and IoT infrastructure, vendor lock-in, the need to deploy and adapt serverless functions as well as their supporting services and software stacks into their cyber-physical execution environment. As a first step towards addressing these challenges, we introduce the SERVERLEss4I0T platform for the design, deployment, and maintenance of applications over the Cloud-Edge-IoT continuum. In particular, our platform enables the specification and deployment of serverless functions on Cloud and Edge resources, as well as the deployment of their supporting services and software stacks over the whole Cloud-Edge-IoT continuum.

The role of Machine Learning (ML) engineers in mobile development has become increasingly important in recent years, as more and more business-critical mobile applications depend on AI components. Many development teams already include dedicated ML engineers who aim to follow agile development practices in their work, as part of the larger MLOps concept. However, the availability of MLOps tools tailored specifically towards mobile platforms is scarce, often due the limited support for non-native programming languages such as Python, as well as the unsuitability of native programming languages such as Java and Kotlin to support ML-related programming tasks. This paper aims to address this gap and describes a plug-in architecture for developing, deploying and running data ingestion and processing components written in Python on the Android platform. With the possibility to pass a user-defined schema with the data format and structure, the proposed architecture ensures that time-series datasets are correctly interpreted by multiple ML modules dealing with both data ingestion and processing,. The proposed approach benefits from modularity, extensibility, customisation, and separation of concerns, which enable ML engineers to be fully involved in a mobile development lifecycle following agile MLOps practices.

Abstract With the increasing computing and networking capabilities, IoT devices and edge gateways have become part of a larger IoT–edge–cloud computing continuum, where processing and storage tasks are distributed across the whole network hierarchy, not concentrated only in the cloud. At the same time, this also introduced continuous delivery practices to the development of software components for network-connected gateways and sensing/actuating nodes. These devices are placed on end users’ premises and are characterized by continuously changing cyber-physical contexts, forcing software developers to maintain multiple application versions and frequently redeploy them on a distributed fleet of devices with respect to their current contexts. Doing this correctly and efficiently goes beyond manual capabilities and requires an intelligent and reliable automated solution. This paper describes a model-based approach to automatically assigning multiple software deployment plans to hundreds of edge gateways and connected IoT devices implemented in collaboration with a smart healthcare application provider. From a platform-specific model of an existing edge computing platform, we extract a platform-independent model that describes a list of target devices and a pool of available deployment plans. Next, we use constraint solving to automatically assign deployment plans to devices at once with respect to their specific contexts. The result is transformed back into the platform-specific model and includes a suitable deployment plan for each device, which is then consumed by our engine to deploy software components not only on edge gateways but also on their downstream IoT devices with constrained resources and connectivity. We validate the approach with a fleet deployment prototype integrated into a DevOps toolchain used by the partner application provider. Initial experiments demonstrate the viability of the approach and its usefulness in supporting DevOps for edge and IoT software development.

The Internet of Things continuously generates avalanches of raw sensor data to be transferred to the Cloud for processing and storage. Due to network latency and limited bandwidth, this vertical offloading model, however, fails to meet requirements of time-critical data-intensive applications which must act upon generated data with minimum time delays. To address such a limitation, this article proposes a novel distributed architecture enabling stream data processing at the edge of the network, broadening the principle of enabling processing closer to data sources adopted by Fog and Edge Computing. Specifically, this architecture extends the Apache NiFi stream processing middleware with support for run-time clustering of heterogeneous edge devices, such that computational tasks can be horizontally offloaded to peer devices and executed in parallel. As opposed to vertical offloading on the Cloud, the proposed solution does not suffer from increased network latency and is thus able to offer 5-25 times faster response time, as demonstrated by the experiments on a run-time license plate recognition system.

Billions of sensor-enabled IoT devices generate extreme amounts of e- waste. Because of the low cost and short lifespan of electronic components, it is often more convenient for consumers to buy a new device instead of re-using or re-purposing the old one. With the increased computing and connectivity capabilities, IoT devices can already receive code updates for new purposes (and thus extend their lifespan), but the cost of such operations often exceeds the price of device replacement due to constrained resources, hindered network connectivity, and distributed placement. This paper describes how these existing capabilities can enable last-mile software deployment at scale. We propose a hierarchical architecture for provisioning software updates from the cloud to terminal devices via edge gateways in a scalable and targeted manner. By enabling such an end-to-end software deployment architecture, the approach promotes hardware re-use via re-purposing and thus contributes to the creation of a more sustainable IoT.

Summary The Internet of Things is underpinned by the global penetration of network‐connected smart devices continuously generating extreme amounts of raw data to be processed in a timely manner. Supported by Cloud and Fog/Edge infrastructures – on the one hand, and Big Data processing techniques – on the other, existing approaches, however, primarily adopt a vertical offloading model that is heavily dependent on the underlying network bandwidth. That is, (constrained) network communication remains the main limitation to achieve truly agile IoT data management and processing. This paper aims to bridge this gap by defining Clustered Edge Computing – a new approach to enable rapid data processing at the very edge of the IoT network by clustering edge devices into fully functional decentralized ensembles, capable of workload distribution and balancing to accomplish relatively complex computational tasks. This paper also proposes ECStream Processing that implements Clustered Edge Computing using Stream Processing techniques to enable dynamic in‐memory computation close to the data source. By spreading the workload among a cluster of collocated edge devices to process data in parallel, the proposed approach aims to improve performance, thereby supporting agile data management. The experimental results confirm that such a distributed in‐memory approach to data processing at the very edge of an IoT network can outperform currently adopted Cloud‐enabled architectures, and has the potential to address a wide range of IoT‐related data‐intensive time‐critical scenarios.

Enabling data processing at the network edge, as close to the actual source of data as possible, is a challenging, yet realistic goal to be achieved by the Internet of Things (IoT), which still primarily relies on the Cloud for data processing. By further extending the Fog and Edge computing principles, recent research advancements enabled aggregation of computing resources from multiple edge devices to support data-intensive task processing using Big Data clustering middleware. The use of these existing solutions, however, is hindered by the heterogeneous, dynamic, mobile, resource-constrained, and time-critical nature of IoT ecosystems. More specifically, a particularly challenging goal is to discover, select, and cluster suitable edge devices - on the one hand, and decompose and allocate data-intensive tasks with respect to discovered resources - on the other. To address this challenge, this paper introduces a novel decentralized architecture for clustering heterogeneous edge devices and executing data-intensive IoT workflows. The proposed approach first breaks down a complex workflow into simpler tasks, then discovers and selects suitable edge devices, and finally allocates the tasks to the selected nodes, connecting them to recompose the original workflow. The proposed approach benefits from an intelligent mapping algorithm that takes into account available cluster resources and processing demands to efficiently allocate fine-grained tasks to selected nodes. To support the clusterisation process, the proposed solution relies on a unified semantic knowledge base that provides a common vocabulary of terms for modelling task requirements and edge device properties, as well as enables automated task grouping and match-making for device discovery and selection, using built-in reasoning capabilities.

Edge computing brings software in close proximity to end users and IoT devices. Given the increasing number of distributed Edge devices with various contexts, as well as the widely adopted continuous delivery practices, software developers need to maintain multiple application versions and frequently (re-)deploy them to a fleet of many devices with respect to their contexts. Doing this correctly and efficiently goes beyond manual capabilities and requires employing an intelligent and reliable automated approach. Accordingly this paper describes a joint research with a Smart Healthcare application provider on a model-based approach to automatically assigning multiple software deployments to hundreds of Edge gateways. From a Platform-Specific Model obtained from the existing Edge computing platform, we extract a Platform-Independent Model that describes a list of target devices and a pool of available deployments. Next, we use constraint solving to automatically assign deployments to devices at once, given their specific contexts. The resulting solution is transformed back to the PSM as to proceed with software deployment accordingly. We validate the approach with a Fleet Deployment prototype integrated into the DevOps toolchain currently used by the application provider. Initial experiments demonstrate the viability of the approach and its usefulness in supporting DevOps in Edge computing applications.

To enable and support smart environments, a recent ICT trend promotes pushing computation from the remote Cloud as close to data sources as possible, resulting in the emergence of the Fog and Edge computing paradigms. Together with Cloud computing, they represent a stacked architecture, in which raw datasets are first pre-processed locally at the Edge and then vertically offloaded to the Fog and/or the Cloud. However, as hardware is becoming increasingly powerful, Edge devices are seen as candidates for offering data processing capabilities, able to pool and share computing resources to achieve better performance at a lower network latency—a pattern that can be also applied to Fog nodes. In these circumstances, it is important to enable efficient, intelligent, and balanced allocation of resources, as well as their further orchestration, in an elastic and transparent manner. To address such a requirement, this article proposes an OpenStack-based middleware platform through which resource containers at the Edge, Fog, and Cloud levels can be discovered, combined, and provisioned to end users and applications, thereby facilitating and orchestrating offloading processes. As demonstrated through a proof of concept on an intelligent surveillance system, by converging the Edge, Fog, and Cloud, the proposed architecture has the potential to enable faster data processing, as compared to processing at the Edge, Fog, or Cloud levels separately. This also allows architects to combine different offloading patterns in a flexible and fine-grained manner, thus providing new workload engineering patterns. Measurements demonstrated the effectiveness of such patterns, even outperforming edge clusters.

The cover image, by Rustem Dautov et al., is based on the Research Article Metropolitan Intelligent Surveillance Systems for Urban Areas by Harnessing IoT and Edge Computing Paradigms, https://doi.org/10.1002/spe.2586. Photo Credit: Rustem Dautov.

Summary Recent technological advances led to the rapid and uncontrolled proliferation of intelligent surveillance systems (ISSs), serving to supervise urban areas. Driven by pressing public safety and security requirements, modern cities are being transformed into tangled cyber‐physical environments, consisting of numerous heterogeneous ISSs under different administrative domains with low or no capabilities for reuse and interaction. This isolated pattern renders itself unsustainable in city‐wide scenarios that typically require to aggregate, manage, and process multiple video streams continuously generated by distributed ISS sources. A coordinated approach is therefore required to enable an interoperable ISS for metropolitan areas, facilitating technological sustainability to prevent network bandwidth saturation. To meet these requirements, this paper combines several approaches and technologies, namely the Internet of Things, cloud computing, edge computing and big data, into a common framework to enable a unified approach to implementing an ISS at an urban scale, thus paving the way for the metropolitan intelligent surveillance system (MISS). The proposed solution aims to push data management and processing tasks as close to data sources as possible, thus increasing performance and security levels that are usually critical to surveillance systems. To demonstrate the feasibility and the effectiveness of this approach, the paper presents a case study based on a distributed ISS scenario in a crowded urban area, implemented on clustered edge devices that are able to off‐load tasks in a “horizontal” manner in the context of the developed MISS framework. As demonstrated by the initial experiments, the MISS prototype is able to obtain face recognition results 8 times faster compared with the traditional off‐loading pattern, where processing tasks are pushed “vertically” to the cloud.

In the era of digital economies, data can be considered as the new commodity, fueling the next-generation software services and applications. Increasing amounts of data, generated on a daily basis by various domains, such as social networks, stock exchanges, the Internet of Things, and cyber-physical systems, are soon expected to exceed the yottabyte1frontier. To process this overwhelming amount, Big Data solutions are being developed to enable a new generation of data-centric/data-intensive applications (DIAs) and services. However, many of such applications currently fail to meet the increasingly demanding data management requirements. In particular, proper techniques and tools to support architects and developers in DIA design are required to cope with these pressing Big Data challenges. This paper makes an initial step in this direction, aiming at reducing the gap between the architects and DIAs they have to develop. The proposed approach extends the conventional Big Data process workflow with a way of capturing and modeling the `three Vs' of Big Data (i.e. volume, velocity, and variety) to provide useful insights on the overall process, knowing the behavior of its individual components. Starting from the V-attributes of the Big Data process components, the proposed framework provides an estimation of its V-metrics by evaluating a performance model generated from the process. To demonstrate the feasibility and the effectiveness of the approach, a case study on a computer vision DIA is reported.