Speaker: Tomasz Nurkiewicz

Loom’s promise: simplify concurrency the way garbage collection simplified memory management. It became transparent and almost forgettable. It's an effort to bring lightweight threads to the JVM. Such threads have low memory and scheduling footprint so that you can create millions of them. Start new threads per each request or connection. Without worrying about memory footprint or context switching. Literally millions. Problems that Loom tries to address:

• obscure non-blocking APIs
• mysterious stack traces coming from reactive frameworks
• pooling threads altogether

That's right; you may no longer need any ExecutorServices. To be honest, it's a leaky abstraction anyway that needs to be fine-tuned and monitored. What if you could simply spawn a new thread whenever it makes sense? Start thinking about tasks to perform, not physical threads that you must manage. Come and see how this radical shift can simplify your concurrent code.

In this talk, we'll explore:

• what is Loom?
• How it's implemented?
• What problems does it solve?
• What are the shortcomings?
• Will it make reactive programming obsolete?

And finally, we'll learn that it's not a silver bullet, and it comes with its own set of issues.