Scaling Web Services for Free to Play
summary
Shows a concrete stack and numbers: customer support runs on Lambda+API Gateway with bug reports auto-posted to Slack and metadata (version, level, receipts) captured; leaderboards write enriched snapshots to S3 and CDN, use local caching (TTL ≈1 min), split into many small 50‑player buckets and validate only high or anomalous scores to reduce cheating. Emphasizes cost tradeoffs with $3.50 per million API Gateway, the cheap alternative CloudFlare workers 50c per million, load‑tested with 24 million sessions over 60 hours and production kill switches and alerts.
AI-written summary · about
chapters (20)
- 0:00Start
- 2:30Free-to-Play Constraints
- 4:09Scaling
- 10:42Serverless
- 14:55Platform-as-A-Service
- 16:22Elastic Beanstalk
- 16:58Kubernetes
- 21:36Security
- 25:13Motivations for Cheating
- 26:53Tools
- 28:34Load Testing
- 29:52Kill Switches
- 30:56Remote Config Tool
- 31:21Database Hotkeys
- 32:39Monitoring and Alerts
- 33:56Customer Support and Leaderboards
- 34:07Customer Support
- 34:31Get Request
- 36:46Customer Support Tool
- 43:44Local Caching
related talks
24mServerless API Gateways for High Scale Systems Performance Cost and Security
54mServerless beyond Functions
47mServerless Microservices
54mYou should be building Serverless Microservices
49mChoosing The Best AWS Service For Your Website + API
20mUnlocking the power API Management for developers
16mBuilding Your Web Application
30mEvolving Serverless Architectures