Wywiad Andrzeja Grzesika z Christin Gorman
summary
Argues that JVM thread-per-request hits memory limits and doesn’t scale, so using Vert.x yields much higher CPU utilization but forces callback-heavy code; points to the Erlang VM and Erlang/Elixir as platforms that support millions of lightweight processes and simpler reasoning about concurrency. Recommends choosing a platform (Go, Erlang/Elixir, or libraries) when high concurrency is required, and calls for JVM-level green-thread support or frameworks that let developers write blocking-style, easy-to-follow distributed code.
AI-written summary · about
chapters (8)
- 0:00Introduction and speaker intro
- 1:04Java fashion and cargo cult
- 3:33Concurrency in Java vs Erlang
- 8:07Alternatives to Java concurrency
- 10:19Encouragement to try new approaches
- 11:25Future of programming environments
- 13:54Challenges of distributed systems
- 15:38Monolithic services vs microservices
related talks
51mWriting Highly Concurrent Polyglot Applications with Vert.x
42mBuilding Bulletproof Concurrent Applications on the JVM
43mConcurrency Oriented Programming in a Modern World
39mLet's talk concurrency
46mJava 21: The revolution of virtual threads - A Deep dive
01h 56mConcurrency Crossroads: Choosing between Reactive Programming & Virtual Threads
46mJava 21: The Revolution of Virtual Threads - A Deep Dive
42mThreading versus Evented