Development of semantic middleware for South Africa’s multi-hazard early warning system

Abstract

The impact of man-made activities and climate change has exacerbated the occurrence of environmental hazards. To improve monitoring systems, integrating Internet of Things (IoT) sensors, legacy systems, and enterprise networks has become crucial for efficient monitoring systems. However, existing heterogenous monitoring systems face issues like data incompatibility, lack of integration, and interoperability, making it challenging for communities to comprehend crucial information. Monitoring systems requirements vary significantly based on the environment, resulting in ad-hoc implementations and integration of different systems and applications. This study addresses these challenges by exploring the use of semantic middleware to harmonize heterogeneous systems, facilitating seamless integration and interoperability within a Multi-Hazard Early System (MHEWS). The study proposes that semantic middleware can overcome the challenges of data representation, integration and system interoperability within a MHEWS. The study focuses on the environmental challenges in mining operations at Lejweleputswa, Free State Province, South Africa as a case study, and adopts two existing Early Warning Systems (EWS), namely Information Technology and Indigenous Knowledge with Intelligence (ITIKI), and Adaptive Environmental Management System (AEMS) frameworks for a MHEWS. The use of semantic representation mechanisms is crucial for connecting diverse systems and facilitating effective communication in a heterogeneous system. This research study demonstrates seamless integration attainability through application containerization, specifically the use of Docker and Kubernetes, to improve the integration, interoperability, deployment, and portability of diverse systems in the context of semantic representation through a middleware. The application of container orchestration platform - Dockers and Kubernetes further strengthens the MHEWS by enabling autoscaling, load balancing, and fault tolerance to ensure high availability even during infrastructure failures. This study achieved the data integration and systems interoperability through the semantic middleware that improves the overall performance of the Multi-Hazard Early Warning System using the case study of Lejweleputswa district, Free State Province, South Africa.