Duke on CRaC with Jakarta EE
summary
Demonstration shows Coordinated Restore at Checkpoint (CRaC) cutting Jakarta EE app cold-starts by roughly 70–75% for the Helidon example, from ~1.4s to ~0.37s, while Quarkus shows little gain; uses Zulu JVM with Zulu with warp engine, checkpoint files written to disk, Docker build scripts, and simple REST app that caches Fibonacci(42) to measure real startup work. Implementation uses the CRaC Resource callbacks beforeCheckpoint and afterRestore plus CDI lifecycle hooks to close/reopen resources.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:26Jakarta EE overview
- 7:25CRaC concept explained
- 13:51CRaC with Jakarta EE
- 27:14Demo setup
- 33:45Running the demo
- 43:10Jakarta EE 12 plans
- 48:29Q&A
related talks
01h 03mDuke on #CRaC with #JakartaEE
53mJava on CRaC: Superfast JVM Application Startup
51mJava on CRaC Superfast JVM Application Startup
53mWhat the CRaC Superfast JVM startup
40mKotlin on CRaC – new JVM features to speed things up
46mCRAC vs GraalVM
45mOptimizing Java Startup and Warmup: Exploring the Power of CRaC on Embedded
47mFast JVM Startup with Checkpoint & Restore