An event-driven fusion framework with auto-scaling of edge intelligence for resilient smart applications in developing countries

Internet of Things has been a popular technology in recent years for deploying a large-scale, smart environment monitoring application across the country, using fog computing and the cloud. However, most locations of the developing countries suffer from power outages and limited network connectivity. Moreover, varied population in different locations, may lead to either frequent or rare changes in the state of the monitored environment. Due to these stochastic conditions, there may be substantial increase in service time of the application with unnecessary battery and resource consumption of the fog node. For efficient utilization of fog node resources in dynamic environment conditions, an event-driven information fusion framework is proposed using docker containerization technology and MQTT (Message Queuing Telemetry Transport) as application layer protocol. The proposed framework provides resilience to the application in the presence of stochastic conditions and auto-scales edge intelligence with minimum scaling operations. The performance of the framework is validated on a smart sanitation system use-case application, proposed for autonomous monitoring of restrooms deployed in Indian rural and semi-urban environment, and the experimental result show deviation of 2.8% in average response time of the application with average accuracy of 98.9%.