Modernize APIs to run serverless using Apache CXF
summary
Takes an existing WSDL, runs Apache CXF codegen to produce Java interfaces and DTOs, packages them with Quarkus, deploys via AWS SAM to Lambda + API Gateway, and demonstrates end-to-end requests working. Cold starts take ~3s on the JVM, enabling Lambda SnapStart drops that to about 1s by snapshotting the Firecracker micro-VM at deploy time; building a GraalVM native image raises build time (~90s) but yields ~0.5s init and reduces memory from ~1GB to ~28MB. Shows concrete configs and trade-offs for WSDL-first codegen.
AI-written summary · about
chapters (14)
- 0:00Introduction and speaker intro
- 1:15What is Apache CXF
- 3:50Why use serverless
- 5:18Modernization scenario
- 6:43API first vs code first
- 9:16Demo setup with CXF
- 12:02Deploying to AWS Lambda
- 14:15Testing the API endpoint
- 17:04Understanding Lambda lifecycle
- 18:50Cold starts and warm starts
- 20:14Optimizing with Lambda SnapStart
- 23:15GraalVM native image benefits
- 25:07Performance comparison results
- 27:37Key takeaways and summary
related talks
52mContainerless, Bloatless, YAML-less, Serverless Java #slideless
56mThinking in Serverless Java (Code)
01h 00mAdapting Java for the Serverless world
49mServerless Java Web Applications on AWS Lambda with Micronaut Quarkus & Spring Boot
53mDissecting Serverless Enterprise Java Applications #slideless
53mServerless Java #slideless
49mJava Patterns and Practices for Serverless Applications
01h 03mKeynote: Why The Future of Java in the Cloud is Serverless