Journey to Cloud : Leave No Java Workload Behind
summary
An engineering tour of Java at cloud scale shows concrete tools and results: IBM Semeru Runtimes built on OpenJ9 adds an embedded shared class cache in container images and an Instant On CRIU checkpoint that drops startup from ~5–15s down to ~0.2s or under 1s, while a remote JIT (JIT server / cloud compiler) decouples compilation, smooths memory spikes and cuts peak RSS substantially, and combined changes boost first‑5‑minute ramp‑up throughput without hurting peak throughput.
AI-written summary · about
chapters (12)
- 0:00Introduction to journey
- 2:15IBM Semeru Runtimes overview
- 5:05Problem space for Java workloads
- 10:39Performance metrics explained
- 14:11Startup and ramp-up challenges
- 18:00Startup time optimization
- 21:49Frugal memory use strategies
- 26:46Shared class cache technology
- 31:53Embedding cache in containers
- 42:03Instant On feature demo
- 53:19Cloud compiler and JIT server
- 1:00:20Summary and future directions
related talks
57mCould your Java workloads be missing out on something?
29mThe Next Frontier in Open Source Java Compilers: Just-in-Time Compilation as a Service
55mWhat's new for Java in clouds?
35mInnovation without compromise: better, stronger, faster Java in the Cloud
24mJITServer: Disaggregated Caching JIT Compiler for the JVM in the Cloud
38mAre you ready for Cloud Native Java?
01h 03mKeeping Your Java Hot.
45mCloud Native JVM. Cloud Runtime