This project aims at extending a flood monitoring system, which adopts a centralized cloud-based approach, to create a novel implementation that adopts a distributed approach based on edge/cloud computing.
NGI-ATLANTIC:OC3-290
EdgeFlooding
EdgeFlooding aims at obtaining a novel implementation of the system that adopts an edge/cloud approach, where part of the data analysis, namely the image analysis, is moved at the edge, in proximity of the monitoring sites. The extension of the original implementation, carried out by UNIPI with the support of UMBC, will restructure the architecture of the platform in order to organize its modules into microservices, thus allowing their deployment on different computing nodes. This modular architecture based on microservices will easily allow to carry out different experiment configurations: one edge/cloud configuration where the image analysis services are deployed on edge nodes while the social media and data aggregation services are hosted on the cloud; and one cloud configuration where all the services are deployed on the cloud as in the original implementation.
In order to assess the performance of such system, an extensive experimentation will be carried out by UNIPI on two phases: a first phase involving only the Grid’5000 testbed, and a second phase involving the Virtual Wall testbed and the Fog testbed at UNIPI. The aim is to assess whether a distributed edge/cloud computing approach is feasible for the implementation of future environmental monitoring systems.
Flash flood monitoring systems for just-in-time notification of flooding events will be crucial to secure any city located in prone flood areas. The results will be of interest not only for future flood monitoring systems, but also more widely for the area of environmental monitoring.
Carlo Vallati, (Department of Information Engineering - University of Pisa)
Not available yet
This experiment is currently underway.
Country: Italy United States
NGI Project: NGI atlantic
Status: Early research demo
Category: Data and machine learningMeasurement, monitoring, analysis and abuse handling