Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 15
Encontrada(s) 1 página(s)

e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publicación Artículo científico (JournalArticle)

Data-centric Service-Based Architecture for Edge-Native 6G Network

  • Baldoni, Gabriele
  • Quevedo, José
  • Magalhaes Guimaraes, Carlos Eduardo
  • Oliva Delgado, Antonio de la
  • Corsaro, Angelo
Starting with 5G, mobile networks moved away from the point-to-point model used by previous generations towards a Service-Based Architecture (SBA) focused on a Cloudnative design. While 5G considers the SBA mainly as a central location with a single administrative domain, 6G is expected to become more of a fully-distributed system. This raises the need for flexible service routing capabilities that allow the utilization of the distributed resources available over multiple infrastructures. This paper embraces this vision and explores the utilization of data-centric and dataflow mechanisms for a seamless realization and interaction of composable services in a fully distributed Edgenative 6G architecture. The feasibility and suitability of this vision are showcased through a proof-of-concept prototype, validated over selected 5G workflows u sing d ifferent t echnologies, with results demonstrating the advantages of pursuing data-centric and dataflow approaches., This project has been partially founded by Horizon 2020 DAEMON (grant no. 101017109), Horizon Europe ICOS (grant no. 101070177), Horizon Europe SNS 6G-IA Hexa-XII (grant no. 101095759), and FCT/MCTES through national funds and when applicable co-funded EU funds under the project UIDB/50008/2020-UIDP/50008/2020
e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publicación Tesis doctoral (DoctoralThesis)

End-to-End convergence of a heterogeneous computing and networking ecosystem

  • Baldoni, Gabriele
In the past decade, cloud computing has transformed the IT landscape, thanks to Infrastructure-as-a-Service (IaaS), offering flexible, cost-effective solutions. However, cloud infrastructure relies on centralized data centers and complex networks, leading to challenges in ensuring quality of service for end users. With the increasing number of connected devices and the emergence ofMulti-access Edge Computing (MEC) and Fog computing paradigms, the cloud industry is evolving to address these challenges. Telco operators are also shifting towards Network Functions Virtualization (NFV), converting central offices into small data centers. This transition creates an interdependency between the computing and networking domains, demanding unified control and management. Industries such as Vehicleto-Everything (V2X), and, Industry 4.0, require optimal performance in both domains, necessitating their convergence. Additionally, the convergence of Operational Technology (OT) and Information Technology (IT) in manufacturing has led to interconnected machines and the concept of Digital Twin. This requires real-time and secure operations on computing and networking domains, extending beyond the edge of the network. To address these evolving challenges, a comprehensive, holistic approach has been taken. This thesis considers a decentralized, heterogeneous, and mobile ecosystem, where software and hardware are geographically distributed. Key challenges include adapting virtualization technologies for resource-constrained devices, addressing application and user mobility, designing an architecture for heterogeneous virtualization infrastructure, and implementing distributed management and orchestration., EUROPEHORIZON2020 5G-DIVE The 5G-DIVE project (2019-2021) is anEU-Taiwan project (H2020-ICT-2019-1 grant agreement 859881) targeting end-to-end 5G trials aiming to prove the technical merits and business value proposition of 5G technologies in two vertical pilots, namely Industry 4.0 and Autonomous Drone Scout. EUROPE HORIZON 2020 DAEMON The DAEMON project (2021-2023) is an Horizon H2020 project (H2020-ICT-2020-2 grant agreement 101017109) that deparing from the current hype around AI, DAEMON will set forth a pragmatic approach to NI design. EUROPEHORIZONEUROPE ICOS The ICOS project (2022-2025) is aHorizon Europe project (HORIZONCL4-2021-DATA-01 grant agreement 101070177) that aims at covering the set of challenges coming up when addressing this continuum paradigm, proposing an approach embedding a well-defined set of functionalities, ending up in the definition of an IoT2cloud Operating System (ICOS)., Programa de Doctorado en Ingeniería Telemática por la Universidad Carlos III de Madrid
e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publicación Artículo científico (JournalArticle)

A Dataflow-Oriented Approach for Machine-Learning-Powered Internet of Things Applications

  • Baldoni, Gabriele
  • Teixeira, Rafael
  • Magalhaes Guimaraes, Carlos Eduardo
  • Antunes, Mario
  • Gomes, Diogo
  • Corsaro, Angelo
The rise of the Internet of Things (IoT) has led to an exponential increase in data generated by connected devices. Machine Learning (ML) has emerged as a powerful tool to analyze these data and enable intelligent IoT applications. However, developing and managing ML applications in the decentralized Cloud-to-Things continuum is extremely complex. This paper proposes Zenoh-Flow, a dataflow programming framework that supports the implementation of End-to-End (E2E) ML pipelines in a fully decentralized manner and abstracted from communication aspects. Thus, it simplifies the development and upgrade process of the next-generation ML-powered applications in the IoT domain. The proposed framework was demonstrated using a real-world use case, and the results showcased a significant improvement in overall performance and network usage compared to the original implementation. Additionally, other of its inherent benefits are a significant step towards developing efficient and scalable ML applications in the decentralized IoT ecosystem., This project has been partially founded by Horizon 2020 DAEMON (grant no. 101017109),Horizon Europe ICOS (grant no. 101070177), and partially supported by CT/MCTES through national funds and when applicable co-funded EU funds under the project UIDB/50008/2020-UIDP/50008/2020.
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Service placement in the continuum: A systematic literature review

  • Sajjad, Waseem|||0009-0007-4956-6995
  • Farreras Esclusa, Montserrat|||0000-0003-2027-1919
  • García Almiñana, Jordi|||0000-0002-3515-7150
  • Masip Bruin, Xavier|||0000-0002-4755-556X
Cloud computing plays a crucial role in the Industry 4.0 era, particularly with the rise of Internet of Things (IoT) applications that support domains such as education, healthcare, business, and manufacturing. These applications consist of multiple services with diverse quality of service (QoS) requirements, making their development and deployment complex. While traditional cloud environments provide scalability, they often fail to support latency-sensitive and resource-intensive applications. To overcome these limitations, alternative paradigms such as Cloud–Fog–Edge (CFE), Cloud–Fog (CF), Cloud–Edge (CE), Fog–Edge (FE), and Mobile Edge Computing (MEC) have emerged. These models push computation, storage, and networking closer to end devices, reducing latency and bandwidth usage. However, the heterogeneity, mobility, and dynamic nature of these environments make service placement (a known NP-hard problem) a central challenge. This article presents a systematic literature review of service placement approaches across the compute continuum. Following established SLR methodology, we identified and analyzed 124 peer-reviewed studies published between 2018 and 2024, classifying them by (i) deployment environment, (ii) service placement strategies and algorithms, (iii) adaptability of the solution, (iv) optimization objectives, (v) virtualization/orchestration technologies, (vi) evaluation methodologies, including workloads, testbeds, and simulation tools and (vii) use cases or application types. The novelty of this work lies in providing not only a detailed taxonomy of placement approaches but also this is the first survey that takes all seven aspects into consideration and establishes correlations between them. Our findings reveal that most existing works target smart health applications and favor heuristic-based placement in complex CFE scenarios, while research on scientific and compute-intensive workloads remains limited. We also identify Kubernetes as the most widely used orchestration technology and latency as the dominant optimization metric. Despite significant progress, the field is still maturing, with gaps in real-world validation and adaptive, ML-based placement strategies. By consolidating technical approaches, evaluation practices, and open challenges, this survey offers both researchers and practitioners a structured overview of the state of the art and guidance for advancing service placement in the compute continuum., The work leading to this publication was partially co-funded by the Spanish Ministry of Science and Innovation under grant PID2021-124463OB-I00 funded by MCIN /AEI /10.13039/501100011033 and by the ERDF ‘‘A way of making Europe’’; by the Catalan Government under contract 2021 SGR 00326; and by the EU Horizon programme under grant 101070177 (ICOS project)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

A survey on IoT-edge-cloud continuum systems: Status, challenges, use cases, and open issues

  • Gkonis, Panagiotis
  • Giannopoulos, Anastasios
  • Trakadas, Panagiotis
  • Masip Bruin, Xavier|||0000-0002-4755-556X
  • D'Andria, Francesco
The rapid growth in the number of interconnected devices on the Internet (referred to as the Internet of Things—IoT), along with the huge volume of data that are exchanged and processed, has created a new landscape in network design and operation. Due to the limited battery size and computational capabilities of IoT nodes, data processing usually takes place on external devices. Since latency minimization is a key concept in modern-era networks, edge servers that are in close proximity to IoT nodes gather and process related data, while in some cases data offloading in the cloud might have to take place. The interconnection of a vast number of heterogeneous IoT devices with the edge servers and the cloud, where the IoT, edge, and cloud converge to form a computing continuum, is also known as the IoT-edge-cloud (IEC) continuum. Several key challenges are associated with this new computing systems’ architectural approach, including (i) the design of connection and programming protocols aimed at properly manipulating a huge number of heterogeneous devices over diverse infrastructures; (ii) the design of efficient task offloading algorithms aimed at optimizing services execution; (iii) the support for security and privacy enhancements during data transfer to deal with the existent and even unforeseen attacks and threats landscape; (iv) scalability, flexibility, and reliability guarantees to face the expected mobility for IoT systems; and (v) the design of optimal resource allocation mechanisms to make the most out of the available resources. These challenges will become even more significant towards the new era of sixth-generation (6G) networks, which will be based on the integration of various cutting-edge heterogeneous technologies. Therefore, the goal of this survey paper is to present all recent developments in the field of IEC continuum systems, with respect to the aforementioned deployment challenges. In the same context, potential limitations and future challenges are highlighted as well. Finally, indicative use cases are also presented from an IEC continuum perspective., This work was partially supported by the project «Towards a functional continuum operating system (ICOS)» funded by the European Commission (Project code/Grant Number: 101070177, call for proposal: HORIZON-CL4-2021-DATA-01, funded under HE|HORIZON-RIA\HORIZON-AG). For the UPC author, this work has been also partially supported by the Spanish Ministry of Science and Innovation under grant PID2021-124463OB-I00, which is funded by MCIN/AEI/10.13039/501100011033 and by ERDF A way of making Europe., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

SBNNR: small-size bat-optimized KNN regression

  • Seyghaly, Rasool|||0000-0002-2805-5607
  • García Almiñana, Jordi|||0000-0002-3515-7150
  • Masip Bruin, Xavier|||0000-0002-4755-556X
  • Kuljanin, Jovana|||0000-0002-3380-262X
Small datasets are frequent in some scientific fields. Such datasets are usually created due to the difficulty or cost of producing laboratory and experimental data. On the other hand, researchers are interested in using machine learning methods to analyze this scale of data. For this reason, in some cases, low-performance, overfitting models are developed for small-scale data. As a result, it appears necessary to develop methods for dealing with this type of data. In this research, we provide a new and innovative framework for regression problems with a small sample size. The base of our proposed method is the K-nearest neighbors (KNN) algorithm. For feature selection, instance selection, and hyperparameter tuning, we use the bat optimization algorithm (BA). Generative Adversarial Networks (GANs) are employed to generate synthetic data, effectively addressing the challenges associated with data sparsity. Concurrently, Deep Neural Networks (DNNs), as a deep learning approach, are utilized for feature extraction from both synthetic and real datasets. This hybrid framework integrates KNN, DNN, and GAN as foundational components and is optimized in multiple aspects (features, instances, and hyperparameters) using BA. The outcomes exhibit an enhancement of up to 5% in the coefficient of determination (¿^2 score) using the proposed method compared to the standard KNN method optimized through grid search., This work has been supported by the Spanish Ministry of Science and Innovation under grant PID2021-124463OB-I00, funded by MCIN/AEI/10.13039/501100011033 and by ERDF “A way of making Europe”, by the European Union’s Horizon Europe under the HE ICOS project, Grant Agreement no. 101070177, and by the Catalan Government under contract 2021 SGR 00326. The corresponding author R.S. gratefully acknowledges the Universitat Politècnica de Catalunya and Banco Santander for the financial support of his predoctoral grant FPI-UPC 2021., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

An open-source AI-as-a-service framework for federated, efficient, and drift-robust learning in the continuum edge–cloud

  • Cajas Ordoñez, Sebastián A.
  • Samanta, Jaydeep
  • Suárez Cetrulo, Andrés L.
  • Ghosh, Romila
  • Giannopoulos, Anastasios
  • Barceló Cuerda, Alex|||0000-0002-0877-5738
  • Simón Carbajo, Ricardo
The rise of edge computing has enabled diverse IoT applications while introducing new challenges in latency, resource constraints, and continuous model adaptations across the cloud, edge, and IoT continuum. Addressing these issues requires optimizing the trade-off between efficiency and latency on resource-constrained hardware, fostering collaboration, and advancing model compression and green computing to reduce computational overhead. We introduce OASIS, an open-source, library-agnostic framework for scalable Edge Machine Learning that unifies predictive analytics, model compression, and supports FL for privacy-preserving training in distributed environments. Designed for real-time forecasting and adaptive monitoring, OASIS enables lightweight ML deployments in dynamic settings with limited resources. OASIS simplifies adoption by offering modular APIs and pre-integrated tools, allowing users to plug in models or connect telemetry data with minimal configuration, making it suitable for practitioners across domains. Our implementation integrates drift detection, SHAP-based explainability, and end-to-end MLOps and monitoring via MLFlow and NannyML. We have presented illustrative examples using real and synthetic data, particularly synthetic CPU telemetry, to stress-test the robustness of the system and demonstrate improvements in inference speed, memory efficiency, and fault resilience. By consolidating critical AI capabilities into a single interface, OASIS lowers the barrier for deploying robust, adaptive, and FL applications at the edge. The code is publicly available here: https://github.com/icos-project/intelligence-module, This work was supported by the HORIZON Research and Innovation Program of European Union under Grant 101070177., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Hierarchical management of extreme-scale task-based applications

  • Lordan Gomis, Francesc|||0000-0002-9845-8890
  • Puigdemunt Schmolling, Gabriel
  • Vergés Boncompte, Pere
  • Conejero Bañón, Francisco Javier
  • Ejarque Artigas, Jorge
  • Badia Sala, Rosa Maria|||0000-0003-2941-5499
The scale and heterogeneity of exascale systems increment the complexity of programming applications exploiting them. Task-based approaches with support for nested tasks are a good-fitting model for them because of the flexibility lying in the task concept. Resembling the hierarchical organization of the hardware, this paper proposes establishing a hierarchy in the application workflow for mapping coarse-grain tasks to the broader hardware components and finer-grain tasks to the lowest levels of the resource hierarchy to benefit from lower-latency and higher-bandwidth communications and exploiting locality. Building on a proposed mechanism to encapsulate within the task the management of its finer-grain parallelism, the paper presents a hierarchical peer-to-peer engine orchestrating the execution of workflow hierarchies with fully-decentralized management. The tests conducted on the MareNostrum 4 supercomputer using a prototype implementation prove the validity of the proposal supporting the execution of up to 707,653 tasks using 2,400 cores and achieving speedups of up to 106 times faster than executions of a single workflow and centralized management., This work has been supported by the Spanish Government (PID2019-107255GB), by MCIN/AEI /10.13039/501100011033 (CEX2021-001148-S), by the Departament de Recerca i Universitats de la Generalitat de Catalunya to the Research Group MPiEDist (2021 SGR 00412), and by the European Commission through the Horizon Europe Research and Innovation program under Grant Agreements 101070177 (ICOS project) and 101016577 (AI-Sprint project)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Offloading artificial intelligence workloads across the computing continuum by means of active storage systems

  • Barceló Cuerda, Alex|||0000-0002-0877-5738
  • Cajas Ordoñez, Sebastián A.
  • Samanta, Jaydeep
  • Suárez Cetrulo, Andrés L.
  • Ghosh, Romila
  • Simón Carbajo, Ricardo
  • Queralt Calafat, Anna|||0000-0003-2782-2955
The increasing demand for artificial intelligence (AI) workloads across diverse computing environments has driven the need for more efficient data management strategies. Traditional cloud-based architectures struggle to handle the sheer volume and velocity of AI-driven data, leading to inefficiencies in storage, computation, and data movement. This paper explores the integration of active storage systems within the computing continuum to optimize AI workload distribution. By embedding computation directly into storage architectures, active storage is able to reduce data transfer overhead, enhancing performance and improving resource utilization. Other existing frameworks and architectures offer mechanisms to distribute certain AI processes across distributed environments; however, they lack the flexibility and adaptability that the continuum requires, both regarding the heterogeneity of devices and the rapid-changing algorithms and models being used by domain experts and researchers. This article proposes a software architecture aimed at seamlessly distributing AI workloads across the computing continuum, and presents its implementation using mainstream Python libraries and dataClay, an active storage platform. The evaluation shows the benefits and trade-offs regarding memory consumption, storage requirements, training times, and execution efficiency across different devices. Experimental results demonstrate that the process of offloading workloads through active storage significantly improves memory efficiency and training speeds while maintaining accuracy. Our findings highlight the potential of active storage to revolutionize AI workload management, making distributed AI deployments more scalable and resource-efficient with a very low entry barrier for domain experts and application developers., This work was partially supported by the projects PID2019107255GB-C21 and PID2023-147979NB-C21, funded by MCIN/AEI/ 10.13039 / 501100011033, and by FEDER, UE. This work was partially supported by the project “Towards a functional continuum operating system (ICOS)” funded by the European Commission under Project code/Grant Number 101070177 through the HORIZON EU program. This work was partially supported by the project “Open CloudEdgeIoT Platform Uptake in Large Scale Cross-Domain Pilots (O-CEI)” funded by the European Commission under Project code/Grant Number 101189589 through the HORIZON EUprogram. Anna Queralt is a Serra-Hunter Fellow and has been partially supported by the Spanish Ministerio de Ciencia e Innovación under project / funding scheme PID2023-152841OAI00 / AEI/10.13039/501100011033 (TALC)., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Offloading artificial intelligence workloads across the computing continuum by means of active storage systems

  • Barceló Cuerda, Alex
  • Cajas Ordoñez, Sebastián A.
  • Samanta, Jaydeep
  • Suárez Cetrulo, Andrés L.
  • Ghosh, Romila
  • Simón Carbajo, Ricardo
  • Queralt Calafat, Anna
The increasing demand for artificial intelligence (AI) workloads across diverse computing environments has driven the need for more efficient data management strategies. Traditional cloud-based architectures struggle to handle the sheer volume and velocity of AI-driven data, leading to inefficiencies in storage, computation, and data movement. This paper explores the integration of active storage systems within the computing continuum to optimize AI workload distribution. By embedding computation directly into storage architectures, active storage is able to reduce data transfer overhead, enhancing performance and improving resource utilization. Other existing frameworks and architectures offer mechanisms to distribute certain AI processes across distributed environments; however, they lack the flexibility and adaptability that the continuum requires, both regarding the heterogeneity of devices and the rapid-changing algorithms and models being used by domain experts and researchers. This article proposes a software architecture aimed at seamlessly distributing AI workloads across the computing continuum, and presents its implementation using mainstream Python libraries and dataClay, an active storage platform. The evaluation shows the benefits and trade-offs regarding memory consumption, storage requirements, training times, and execution efficiency across different devices. Experimental results demonstrate that the process of offloading workloads through active storage significantly improves memory efficiency and training speeds while maintaining accuracy. Our findings highlight the potential of active storage to revolutionize AI workload management, making distributed AI deployments more scalable and resource-efficient with a very low entry barrier for domain experts and application developers., This work was partially supported by the projects PID2019107255GB-C21 and PID2023-147979NB-C21, funded by MCIN/AEI/ 10.13039 / 501100011033, and by FEDER, UE. This work was partially supported by the project “Towards a functional continuum operating system (ICOS)” funded by the European Commission under Project code/Grant Number 101070177 through the HORIZON EU program. This work was partially supported by the project “Open CloudEdgeIoT Platform Uptake in Large Scale Cross-Domain Pilots (O-CEI)” funded by the European Commission under Project code/Grant Number 101189589 through the HORIZON EUprogram. Anna Queralt is a Serra-Hunter Fellow and has been partially supported by the Spanish Ministerio de Ciencia e Innovación under project / funding scheme PID2023-152841OAI00 / AEI/10.13039/501100011033 (TALC)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Hierarchical management of extreme-scale task-based applications

  • Lordan Gomis, Francesc
  • Puigdemunt Schmolling, Gabriel
  • Vergés Boncompte, Pere
  • Conejero Bañón, Francisco Javier
  • Ejarque Artigas, Jorge
  • Badia Sala, Rosa Maria
The scale and heterogeneity of exascale systems increment the complexity of programming applications exploiting them. Task-based approaches with support for nested tasks are a good-fitting model for them because of the flexibility lying in the task concept. Resembling the hierarchical organization of the hardware, this paper proposes establishing a hierarchy in the application workflow for mapping coarse-grain tasks to the broader hardware components and finer-grain tasks to the lowest levels of the resource hierarchy to benefit from lower-latency and higher-bandwidth communications and exploiting locality. Building on a proposed mechanism to encapsulate within the task the management of its finer-grain parallelism, the paper presents a hierarchical peer-to-peer engine orchestrating the execution of workflow hierarchies with fully-decentralized management. The tests conducted on the MareNostrum 4 supercomputer using a prototype implementation prove the validity of the proposal supporting the execution of up to 707,653 tasks using 2,400 cores and achieving speedups of up to 106 times faster than executions of a single workflow and centralized management., This work has been supported by the Spanish Government (PID2019-107255GB), by MCIN/AEI /10.13039/501100011033 (CEX2021-001148-S), by the Departament de Recerca i Universitats de la Generalitat de Catalunya to the Research Group MPiEDist (2021 SGR 00412), and by the European Commission through the Horizon Europe Research and Innovation program under Grant Agreements 101070177 (ICOS project) and 101016577 (AI-Sprint project)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

SBNNR: small-size bat-optimized KNN regression

  • Seyghaly, Rasool
  • García Almiñana, Jordi
  • Masip Bruin, Xavier
  • Kuljanin, Jovana
Small datasets are frequent in some scientific fields. Such datasets are usually created due to the difficulty or cost of producing laboratory and experimental data. On the other hand, researchers are interested in using machine learning methods to analyze this scale of data. For this reason, in some cases, low-performance, overfitting models are developed for small-scale data. As a result, it appears necessary to develop methods for dealing with this type of data. In this research, we provide a new and innovative framework for regression problems with a small sample size. The base of our proposed method is the K-nearest neighbors (KNN) algorithm. For feature selection, instance selection, and hyperparameter tuning, we use the bat optimization algorithm (BA). Generative Adversarial Networks (GANs) are employed to generate synthetic data, effectively addressing the challenges associated with data sparsity. Concurrently, Deep Neural Networks (DNNs), as a deep learning approach, are utilized for feature extraction from both synthetic and real datasets. This hybrid framework integrates KNN, DNN, and GAN as foundational components and is optimized in multiple aspects (features, instances, and hyperparameters) using BA. The outcomes exhibit an enhancement of up to 5% in the coefficient of determination (¿^2 score) using the proposed method compared to the standard KNN method optimized through grid search., This work has been supported by the Spanish Ministry of Science and Innovation under grant PID2021-124463OB-I00, funded by MCIN/AEI/10.13039/501100011033 and by ERDF “A way of making Europe”, by the European Union’s Horizon Europe under the HE ICOS project, Grant Agreement no. 101070177, and by the Catalan Government under contract 2021 SGR 00326. The corresponding author R.S. gratefully acknowledges the Universitat Politècnica de Catalunya and Banco Santander for the financial support of his predoctoral grant FPI-UPC 2021., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

An open-source AI-as-a-service framework for federated, efficient, and drift-robust learning in the continuum edge–cloud

  • Cajas Ordoñez, Sebastián A.
  • Samanta, Jaydeep
  • Suárez Cetrulo, Andrés L.
  • Ghosh, Romila
  • Giannopoulos, Anastasios
  • Barceló Cuerda, Alex
  • Simón Carbajo, Ricardo
The rise of edge computing has enabled diverse IoT applications while introducing new challenges in latency, resource constraints, and continuous model adaptations across the cloud, edge, and IoT continuum. Addressing these issues requires optimizing the trade-off between efficiency and latency on resource-constrained hardware, fostering collaboration, and advancing model compression and green computing to reduce computational overhead. We introduce OASIS, an open-source, library-agnostic framework for scalable Edge Machine Learning that unifies predictive analytics, model compression, and supports FL for privacy-preserving training in distributed environments. Designed for real-time forecasting and adaptive monitoring, OASIS enables lightweight ML deployments in dynamic settings with limited resources. OASIS simplifies adoption by offering modular APIs and pre-integrated tools, allowing users to plug in models or connect telemetry data with minimal configuration, making it suitable for practitioners across domains. Our implementation integrates drift detection, SHAP-based explainability, and end-to-end MLOps and monitoring via MLFlow and NannyML. We have presented illustrative examples using real and synthetic data, particularly synthetic CPU telemetry, to stress-test the robustness of the system and demonstrate improvements in inference speed, memory efficiency, and fault resilience. By consolidating critical AI capabilities into a single interface, OASIS lowers the barrier for deploying robust, adaptive, and FL applications at the edge. The code is publicly available here: https://github.com/icos-project/intelligence-module, This work was supported by the HORIZON Research and Innovation Program of European Union under Grant 101070177., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

A survey on IoT-edge-cloud continuum systems: Status, challenges, use cases, and open issues

  • Gkonis, Panagiotis
  • Giannopoulos, Anastasios
  • Trakadas, Panagiotis
  • Masip Bruin, Xavier
  • D'Andria, Francesco
The rapid growth in the number of interconnected devices on the Internet (referred to as the Internet of Things—IoT), along with the huge volume of data that are exchanged and processed, has created a new landscape in network design and operation. Due to the limited battery size and computational capabilities of IoT nodes, data processing usually takes place on external devices. Since latency minimization is a key concept in modern-era networks, edge servers that are in close proximity to IoT nodes gather and process related data, while in some cases data offloading in the cloud might have to take place. The interconnection of a vast number of heterogeneous IoT devices with the edge servers and the cloud, where the IoT, edge, and cloud converge to form a computing continuum, is also known as the IoT-edge-cloud (IEC) continuum. Several key challenges are associated with this new computing systems’ architectural approach, including (i) the design of connection and programming protocols aimed at properly manipulating a huge number of heterogeneous devices over diverse infrastructures; (ii) the design of efficient task offloading algorithms aimed at optimizing services execution; (iii) the support for security and privacy enhancements during data transfer to deal with the existent and even unforeseen attacks and threats landscape; (iv) scalability, flexibility, and reliability guarantees to face the expected mobility for IoT systems; and (v) the design of optimal resource allocation mechanisms to make the most out of the available resources. These challenges will become even more significant towards the new era of sixth-generation (6G) networks, which will be based on the integration of various cutting-edge heterogeneous technologies. Therefore, the goal of this survey paper is to present all recent developments in the field of IEC continuum systems, with respect to the aforementioned deployment challenges. In the same context, potential limitations and future challenges are highlighted as well. Finally, indicative use cases are also presented from an IEC continuum perspective., This work was partially supported by the project «Towards a functional continuum operating system (ICOS)» funded by the European Commission (Project code/Grant Number: 101070177, call for proposal: HORIZON-CL4-2021-DATA-01, funded under HE|HORIZON-RIA\HORIZON-AG). For the UPC author, this work has been also partially supported by the Spanish Ministry of Science and Innovation under grant PID2021-124463OB-I00, which is funded by MCIN/AEI/10.13039/501100011033 and by ERDF A way of making Europe., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Service placement in the continuum: A systematic literature review

  • Sajjad, Waseem
  • Farreras Esclusa, Montserrat
  • García Almiñana, Jordi
  • Masip Bruin, Xavier
Cloud computing plays a crucial role in the Industry 4.0 era, particularly with the rise of Internet of Things (IoT) applications that support domains such as education, healthcare, business, and manufacturing. These applications consist of multiple services with diverse quality of service (QoS) requirements, making their development and deployment complex. While traditional cloud environments provide scalability, they often fail to support latency-sensitive and resource-intensive applications. To overcome these limitations, alternative paradigms such as Cloud–Fog–Edge (CFE), Cloud–Fog (CF), Cloud–Edge (CE), Fog–Edge (FE), and Mobile Edge Computing (MEC) have emerged. These models push computation, storage, and networking closer to end devices, reducing latency and bandwidth usage. However, the heterogeneity, mobility, and dynamic nature of these environments make service placement (a known NP-hard problem) a central challenge. This article presents a systematic literature review of service placement approaches across the compute continuum. Following established SLR methodology, we identified and analyzed 124 peer-reviewed studies published between 2018 and 2024, classifying them by (i) deployment environment, (ii) service placement strategies and algorithms, (iii) adaptability of the solution, (iv) optimization objectives, (v) virtualization/orchestration technologies, (vi) evaluation methodologies, including workloads, testbeds, and simulation tools and (vii) use cases or application types. The novelty of this work lies in providing not only a detailed taxonomy of placement approaches but also this is the first survey that takes all seven aspects into consideration and establishes correlations between them. Our findings reveal that most existing works target smart health applications and favor heuristic-based placement in complex CFE scenarios, while research on scientific and compute-intensive workloads remains limited. We also identify Kubernetes as the most widely used orchestration technology and latency as the dominant optimization metric. Despite significant progress, the field is still maturing, with gaps in real-world validation and adaptive, ML-based placement strategies. By consolidating technical approaches, evaluation practices, and open challenges, this survey offers both researchers and practitioners a structured overview of the state of the art and guidance for advancing service placement in the compute continuum., The work leading to this publication was partially co-funded by the Spanish Ministry of Science and Innovation under grant PID2021-124463OB-I00 funded by MCIN /AEI /10.13039/501100011033 and by the ERDF ‘‘A way of making Europe’’; by the Catalan Government under contract 2021 SGR 00326; and by the EU Horizon programme under grant 101070177 (ICOS project)., Peer Reviewed, Postprint (published version)