Crushing Application Latency with Vert.x
summary
Demo builds ES4X runtime (~3,000 lines) using Vert.x for I/O and GraalVM for JIT and module loading, and measures against TechEmpower benchmarks. Non‑blocking code shows ~60% higher throughput (example: ~37k vs ~14k req/s), Node.js example jumps from rank 89 to 6, and a mixed Java/JS run gains ~1,500 req/s over a pure Java version. Under heavy concurrent load measured latency falls below the blocking Java implementations.
AI-written summary · about
chapters (10)
- 0:00Introduction and speaker
- 1:50Latency as a system symptom
- 5:47What is Vert.x
- 7:16Problem with blocking I/O
- 14:21Benefits of non-blocking I/O
- 17:08Benchmarking Vert.x vs blocking APIs
- 21:08Polyglot programming challenges
- 25:09GraalVM for polyglot performance
- 27:13es4x: JavaScript for Vert.x
- 31:53Conclusion and takeaways
related talks
51mWriting Highly Concurrent Polyglot Applications with Vert.x
46mReal-world HTTP benchmarking
01h 01mReactive Microservices on the JVM with Vert.x
54mReal-World HTTP Performance Benchmarking, Lessons Learned
38mReactive ❤️ Loom: A Forbidden Love Story
52mGraalVM: Run Programs Faster Anywhere
01h 01mExploring Vert.x in Java, Groovy, and JavaScript
45mReactive web services