Skip to content

Concurrency

How a server handles many requests at once — the models (threads, event loops, async I/O), their tradeoffs, and when each wins. Cross-cutting: the same concepts show up in web servers, GraphQL servers, databases, and language runtimes.

Written

Event loop vs threads
Thread-per-request vs single-threaded async; OS vs user-level threads.
Planned topics
  • Node.js event loop internals — phases (timers, poll, check), the microtask queue, libuv thread pool for fs/DNS, what actually blocks the loop.
  • Project Loom (JVM virtual threads) — thread-per-request coding model on top of a small number of OS carrier threads; how it removes the “blocking wastes a thread” penalty.
  • Goroutines & the Go scheduler — M:N scheduling, user-level threads, work-stealing.
  • I/O-bound vs CPU-bound — the deciding axis for which model wins; how to tell which you have.
  • How Netflix DGS handles concurrency — GraphQL on the JVM (Spring), reactive vs blocking, contrast with Node/Apollo.
  • Reactive / async frameworks — event-driven vs virtual-thread approaches to the same problem.