Funlish. Ready when the whole class signs in.
Funlish, owned and operated by Techlix, helps children learn English through browser-based games, voices and pictures. We moved it from a single server to an AWS platform built around school-hour demand, protected learning progress and media delivery across phones, tablets and computers.
Every new class depended on the same server.
Funlish ran its Node.js application, Nginx web server, SQLite database and media library on one EC2 instance. Teachers and administrators created children's accounts and followed their progress through the same site. An instance or disk failure could interrupt lessons and put that progress at risk.
Whole-class logins competed with media delivery and progress updates for fixed capacity. Techlix needed a platform that could grow with school demand, protect login and administration, and surface problems through monitoring rather than reports from teachers and parents.
Separate the lesson, the media and the learning record.
Move learning progress off the application server
Children's accounts and progress moved from SQLite to Amazon RDS for MySQL with Multi-AZ replication, encryption and automated backups. Migration rehearsals checked record counts and checksums before the final sync.
Deliver games and media closer to learners
Games, voices and pictures moved to a private, versioned S3 bucket behind CloudFront Origin Access Control. Content-specific caching and compression reduce origin load, leaving the application tier to handle login, game logic and progress.
Prepare for whole-class logins
An EC2 Auto Scaling group runs the Node.js application across two zones behind an Application Load Balancer. AWS WAF adds managed rules, bot controls and rate limits for login and administration, tuned before blocking traffic.
Make the platform repeatable and observable
Terraform defines the infrastructure, with Checkov scanning in the delivery pipeline. A proof of concept exercised the real application before production, while CloudWatch dashboards, alerts, runbooks and training support day-to-day operation.
A platform the team can scale and operate.
The delivered platform separates application traffic, learning records and media delivery. Accounts and progress migrated without data loss, and validation covered database failover, backup restoration, whole-class load scenarios and WAF rules. Techlix received the infrastructure code, operations runbooks and training to manage the platform.
The architecture is sized for around 7,000 concurrent learners at school-hours peak. That figure describes designed capacity, not measured production attendance. Auto Scaling adjusts application capacity with demand, while CloudFront serves the media library.