Data-Driven Operational Artificial Intelligence for Computing Continuum: A Natural Disaster Management Use Case

The increasing frequency of documented natural disasters can be attributed to advances in communication technologies, such as satellites, the Internet, and smart devices that facilitate better disaster reporting. This is coupled with an actual rise in the occurrence of such events and improved documentation of their impacts. These trends underscore the pressing need for scalable and intelligent technological solutions to efficiently process large datasets, allowing informed decision-making and effective disaster response. This study presents a Computing Continuum framework that integrates intelligence across cloud, edge and deep edge tiers for efficient disaster data processing. A significant characteristic is the incorporation of Artificial Intelligence for IT Operations (AIOps), which leverages machine learning and analytics to facilitate dynamic resource management and adaptive system modeling, thereby addressing the intricate challenges posed by disaster scenarios. The architecture encompasses an AI-driven framework for monitoring and managing service, network, and infrastructure layers, tailoring policies to specific disaster needs. The proposed framework is applied to wildfire management, leveraging an AI Operation Manager to coordinate sensor-equipped drones for real-time data acquisition and processing. Operating at the deep edge tier, these drones transmit environmental data to edge and cloud infrastructures for analysis. This multi-tiered approach improves situational awareness, disaster response, and resource utilization.