The COGNIFOG project has achieved several significant results. From the early stages, a reference architecture was established, covering application modeling tools, deployment infrastructure configuration tools, DevOps methodologies, runtime tools, and governance tools.
The COGNIFOG framework is a flexible, adaptive, and modular system designed to facilitate the deployment of applications across a cognitive edge-cloud continuum. It offers comprehensive modeling and configuration tools to describe application constraints and requirements. The framework's orchestration and load-balancing capabilities ensure that both functional and non-functional requirements, such as energy cost, latency, and resource availability, are met. Additionally, COGNIFOG provides advanced real-time observability features, allowing for the monitoring of various metrics like CPU load, memory usage, latency, response time, and bandwidth. The framework also enhances privacy, security, and safety, making it a robust solution for diverse application domains. COGNIFOG has been successfully validated through three case studies: flood detection and coordination of rescue missions, telemedicine, and smart manufacturing, demonstrating its effectiveness in real-world scenarios.
The COGNIFOG framework's capabilities were demonstrated through three distinct industrial case studies. In flood detection and rescue coordination (provided by Thales), COGNIFOG's multi-clustering and smart allocation capabilities enabled optimal deployment and management of geographically distributed Edge nodes, reducing resource and energy consumption, and facilitating seamless communication between edge and cloud resources. For telemedicine in isolated areas (provided by TMA), the framework showcased its scalability and reliability, securely handling large volumes of concurrent telemetry data from smartwatches and e-health suitcases, while significantly reducing energy consumption by leveraging lightweight edge servers. Meanwhile, smart manufacturing case study (provided by LMS) highlighted COGNIFOG's capability in managing autonomous robotic operations, with its containerized deployments, Kubernetes orchestration, and dynamic workload allocation that ensures low-latency, efficient, and resilient performance. Each case study tailored the framework to its unique needs, underscoring COGNIFOG's adaptability and effectiveness across diverse domains.
The KPIs measured during the execution of the case studies showed a drastic reduction in energy consumption and CO2 emissions (from 22% in the LMS case study to 75% in the TMA and Thales case studies). TMA demonstrated that COGNIFOG can support scalability, allowing the management of up to 5400 connected watches. In the three case studies, the infrastructure deployment time is less than 8 minutes (including hardware and software deployment); the integration time of an IoT device is less than 1 minute. Moreover, no data loss was observed, demonstrating that data integrity is 100%. These results testify to the effectiveness and robustness of the COGNIFOG framework.
