Achieving High Availability through Implementation of Message Queue on Document Forgery Software

The growth and demand of falsification detection software is in tandem with the amount of data processed, thus driving the need to build a system capable of handling this amount of data. This study aims to introduce Siberpunk - a microservice-based applications using message queues to detect document forgery. This research also uses containers to support high availability. The process of testing the system in this study uses the load testing method by comparing the implementation of message queues and REST on the system being built. Load testing in research uses the help of the Apache JMeter application. The results of this study indicate that the developed system can separate the dependence of the input process on the plagiarism detection system, and the message queue performance results show that RabbitMQ has stable memory consumption and Kafka has lowest latency.