Keynote: Why The Future of Java in the Cloud is Serverless
summary
Argues Java works well serverless by showing a runnable template that builds a Quarkus + AWS Lambda + CDK stack, deploys a MicroProfile JAX‑RS app, and measures cold starts ~3s then warm responses ~5–160ms. Recommends fat functions (one lambda per microservice, not per HTTP verb), uses Corquus/Quarkus build‑time optimization, and shows costs: ~ $30/month for a 2 GB function at 200 ms avg, ~$21/month to keep a function always warm. Notes limits: no JVM-wide shared statics, timers need EventBridge, use S3/EFS for file I/O.
AI-written summary · about
chapters (8)
- 0:00Start
- 3:10Java Web Server
- 4:22Egb Clusters
- 14:44What Is Fn Java
- 26:29Serverless Best Practices with Python
- 29:20Java Function Design
- 40:44Source Code
- 47:59Event Bridge
related talks
56mThinking in Serverless Java (Code)
52mContainerless, Bloatless, YAML-less, Serverless Java #slideless
49mJava Patterns and Practices for Serverless Applications
53mServerless Java #slideless
53mDissecting Serverless Enterprise Java Applications #slideless
53m"Serverless" Is What J2EE Was Meant To Be
48mJava in the Clouds: Pointless and Expensive or Productive and Must-Have?
36mServerless Java