Design of a Geographic Information System for Forest and Land Fires Based on a Real-Time Database on Microservices Infrastructure

Forest and land fires are an increasingly common problem in Indonesia. Fire cases that often occur require a system that is able to detect fires and provide information to users remotely to reduce the impact of fires. Along with the development of hardware technology such as computers, the use of GIS seems to be an effective shortcut in analyzing an event. Kubernetes is an open source platform for managing containerized application workloads, offering declarative configuration and automation. This research designs a Google Maps API System tool for forest and land fires using a real-time database on microservices infrastructure with outputs in the form of fire locations and the results of sensor readings used. Broadly speaking, the processes that occur in the design of the location of forest fire points will be detected by sensors. Then firebase will store forest fire data which will simultaneously be updated on the website. Clients can see the point of forest fires through a browser on their respective desktops. Based on the results of the performance tests that have been carried out, it can be concluded that the use of Kubernetes microclusters can provide advantages when compared to those built monolithically, because Kubernetes microclusters have several advantages, namely having the Horizontal Pod Autoscaler feature, and the Kubernetes microcluster manages components and related services well. Then for each test performed, there was no significant change in memory usage. In the analysis of the results of the comparison data with 7 tests that have been carried out there are 6 tests which mean that the service built with the Kubernetes microcluster is superior to the monolithic one, namely hits per second 2354 ms, latency 3599 ms, response code 720 success code, cpu utilization 13.84%, error rate error rate 0.00%, and throughput 112/sec.