The promise of Reactive programming models is that you can free yourself from the constraints of handling one request for each thread and realize increased throughput as a result. The only problem is that it requires a completely different set of APIs that many developers find counter-intuitive. What if you can achieve the same performance using thread-per-request APIs, and let the Java virtual machine handle the hard work of blocking when appropriate, and executing platform threads when the time is right? Enter virtual threads, a key feature of Project Loom, currently available in JDK 19 and higher. In this session, we'll look at how Java servers and frameworks, such as Tomcat, Jetty, Spring Boot, Helidon and Quarkus, are using this powerful new feature to increase throughput without requiring reactive programming models.

Services Reloaded: Increased Throughput with Project Loom Virtual Threads