Periodic Reporting for period 1 - COGNIFOG (AI-empowered Edge Cloud Continuum for self-aware cognitive computing environments)
Periodo di rendicontazione: 2023-01-01 al 2024-06-30
Cognitive-Fog is the concept enabling the IoT-to-edge-to-cloud continuum. Despite its heterogeneous nature, the fog side allows dynamic resource allocation and sharing at many sub-layers between the cloud and the edge. In the IoT side, it provides interoperability facilities so that IoT devices can be easily connected to the continuum and communicate with other edge-side components. On the Edge side, it can be hosted on microservers, where it operates as an extension of the centralised cloud. On the cloud side, it will help to orchestrate all necessary resources so that an end-to-end service is provided in a safe and secure way and in both directions. Such technology, covering all these aspects, does not exist and needs to be designed and developed.
COGNIFOG provides an efficient tool that will facilitate the orchestration of heterogeneous distributed computing resources in a wide where IoT, Edge and Cloud ecosystem could interact and build a harmonized fog continuum. COGNIFOG framework provides distinct, loosely coupled tools, each addressing specific functionalities vital to the continuum.
COGNIFOG allows the deployment of applications at different levels (far edge, near edge, and cloud). The modelling tools will enable the description of various applications in various domains. The orchestrator allows the workload to be distributed automatically in a Kubernetes multi-cluster environment according to the application's needs. A comprehensive monitoring stack has been developed and integrated, enabling real-time monitoring and management of the system's performance and health. A multi-cluster orchestrator manages applications and services across multiple clusters, spanning on-premises, hybrid, and multi-cloud environments while mitigating complexity, enhancing scalability, ensuring resilience, and bolstering security and compliance. The interoperability is smoothly ensured between cloud and edge levels and between edge and IoT levels. Finally, QoS Solver mainly considers energy efficiency and sustainability to realise energy-efficient placement.
COGNIFOG framework will be validated on three different domains: coordinated crisis management in urban areas (Thales); E-health services in the Edge-Cloud continuum (TMA) and Automated Edge-Cloud continuum for Smart Manufacturing (LMS).
COGNIFOG provides an efficient tool that will facilitate the orchestration of heterogeneous distributed computing resources in a wide where IoT, Edge and Cloud ecosystem could interact and build a harmonized fog continuum. COGNIFOG framework provides distinct, loosely coupled tools, each addressing specific functionalities vital to the continuum.
COGNIFOG allows the deployment of applications at different levels (far edge, near edge, and cloud). The modelling tools will enable the description of various applications in various domains. The orchestrator allows the workload to be distributed automatically in a Kubernetes multi-cluster environment according to the application's needs. A comprehensive monitoring stack has been developed and integrated, enabling real-time monitoring and management of the system's performance and health. A multi-cluster orchestrator manages applications and services across multiple clusters, spanning on-premises, hybrid, and multi-cloud environments while mitigating complexity, enhancing scalability, ensuring resilience, and bolstering security and compliance. The interoperability is smoothly ensured between cloud and edge levels and between edge and IoT levels. Finally, QoS Solver mainly considers energy efficiency and sustainability to realise energy-efficient placement.
COGNIFOG framework will be validated on three different domains: coordinated crisis management in urban areas (Thales); E-health services in the Edge-Cloud continuum (TMA) and Automated Edge-Cloud continuum for Smart Manufacturing (LMS).
The main achievement in this period is the proposition of the COGNIFOG reference architecture, the integration activities to develop the COGNIFOG framework as well as the use cases specifications to either identify the generic applications needs as well as the specific ones.
The COGNIFOG framework is multidimensional and covers the entire workload lifecycle, from design and development to execution and data generation. The framework integrates 23 decoupled modules provided by the partners. The modelling tools enable the description of various applications in various domains. The orchestrator allows the workload to be distributed automatically in a Kubernetes multi-cluster environment according to the application’s needs. The interoperability is ensured by the Slice Manager that effectively collaborate with the orchestrator to manage and orchestrate network slices while integrating the new IoT edge gateway, enabling dynamic and efficient management of network slices at the edge, tailored to the specific needs of IoT devices and applications. A comprehensive monitoring stack has been developed and integrated, enabling real-time monitoring and management of the system’s performance and health. The clever allocator collects monitoring data from the clusters. For workload scheduling, QoS Solver mainly considers energy efficiency and sustainability. To this end, a cost criterion is incorporated into QoS Solver, which should be adapted to realise energy-efficient placement.
The integration activities are still ongoing but first demonstrators and proof of concepts has already shown the feasibility and eliminate technical risks. The integration activities will continue as planned in the GA until January 2025 (M24). The specification activities (WP2) are alsmost finished while validation activities in the trials began in March 2024. Trial Partners have defined the high-level goals of the trials, the hardware and system components that will be used, the data paths exercised in the trials, the scenarios description, the set-up procedures for each scenario. The validation activity will continue until the end of the project.
The COGNIFOG framework is multidimensional and covers the entire workload lifecycle, from design and development to execution and data generation. The framework integrates 23 decoupled modules provided by the partners. The modelling tools enable the description of various applications in various domains. The orchestrator allows the workload to be distributed automatically in a Kubernetes multi-cluster environment according to the application’s needs. The interoperability is ensured by the Slice Manager that effectively collaborate with the orchestrator to manage and orchestrate network slices while integrating the new IoT edge gateway, enabling dynamic and efficient management of network slices at the edge, tailored to the specific needs of IoT devices and applications. A comprehensive monitoring stack has been developed and integrated, enabling real-time monitoring and management of the system’s performance and health. The clever allocator collects monitoring data from the clusters. For workload scheduling, QoS Solver mainly considers energy efficiency and sustainability. To this end, a cost criterion is incorporated into QoS Solver, which should be adapted to realise energy-efficient placement.
The integration activities are still ongoing but first demonstrators and proof of concepts has already shown the feasibility and eliminate technical risks. The integration activities will continue as planned in the GA until January 2025 (M24). The specification activities (WP2) are alsmost finished while validation activities in the trials began in March 2024. Trial Partners have defined the high-level goals of the trials, the hardware and system components that will be used, the data paths exercised in the trials, the scenarios description, the set-up procedures for each scenario. The validation activity will continue until the end of the project.
As said previously, a technology providing an automatic and an easy integration of the different layers in a COGNITIVE-Fog (IoT-to-edge-to-cloud continuum) does not exist yet.
COGNIFOG contributes to the state of the art advancement by integrating different tools covering different functionalities and purpose. For instance, Rancher Fleet (https://fleet.rancher.io/(si apre in una nuova finestra)) and Open Cluster Management(https://open-cluster-management.io/(si apre in una nuova finestra)) has simplified deployment, scaling, resource management, and enhanced security across decentralized infrastructure ranging from cloud, on-premises, edge and far-edge devices built do not include modeling and monitoring tools like COGNIFOG. COGNIFOG also addresses the challenge of multi-level interoperability by adopting the “on-site interoperability” approach within the IoT edge gateway and introduces “IoT data interoperability” using Eclipse sensiNact platform developed by KENTYOU which is a framework that enables data exchange and compatibility across a broader spectrum of IoT devices and services. Finally, to face the security challenges, COGNIFOG builds its security strategy upon two foundational axioms: "Security by Design" and "End-to-End Security". Both principles emphasise that security is integral to the architecture, not just an afterthought.
COGNIFOG contributes to the state of the art advancement by integrating different tools covering different functionalities and purpose. For instance, Rancher Fleet (https://fleet.rancher.io/(si apre in una nuova finestra)) and Open Cluster Management(https://open-cluster-management.io/(si apre in una nuova finestra)) has simplified deployment, scaling, resource management, and enhanced security across decentralized infrastructure ranging from cloud, on-premises, edge and far-edge devices built do not include modeling and monitoring tools like COGNIFOG. COGNIFOG also addresses the challenge of multi-level interoperability by adopting the “on-site interoperability” approach within the IoT edge gateway and introduces “IoT data interoperability” using Eclipse sensiNact platform developed by KENTYOU which is a framework that enables data exchange and compatibility across a broader spectrum of IoT devices and services. Finally, to face the security challenges, COGNIFOG builds its security strategy upon two foundational axioms: "Security by Design" and "End-to-End Security". Both principles emphasise that security is integral to the architecture, not just an afterthought.