Improving Moodle Architecture and Learning Features in Cloud Server Ecosystem Using Kubernetes and Gamification

Abstract

LMS has been used by many universities and schools to improve learning efficiency and effectiveness. One of the many LMS is Moodle. This LMS has many learning features and can be extended by installing “plugins”. However, Moodle is monolith, which means all its resources like front-end, files, and database are installed in the same server and requires high cost server to do so. This study presents Moodle LMS architecture hosted in Google Cloud Platform to solve this problem. The core LMS learning features are extended by adding location-based learning and object-recognition learning as gamification element. Both features are implemented in the form of Moodle plugins that utilizes external service via API. Location-based learning utilizes Leaflet interactive maps and Google Maps place autocomplete search. While object-recognition learning utilizes an object recognition service (ORS) hosted in Google Kubernetes Engine to classify, store, and create image models. Evaluation of the architecture uses the Architecture Trade-off Analysis Method (ATAM) to evaluate performance and scalability as quality attributes. Experimental results showed that Moodle core LMS could support 480 concurrent users with average response time of 11.37 seconds with 100% availability. While ORS could support 960 concurrent users with availability above 98%. Thus, educators could use this architecture as an example to support many students. In the future, this solution may be evaluated if it could increase student motivation and achievement.