Multi-tiered Node.js architectures
summary
Claim: reduce client response time by shrinking the critical path and moving non-user‑facing work off the event loop. Uses cluster to prefork ~20 workers and pass the Port 80 socket at the kernel level, uses file descriptors / Unix sockets for fast IPC, and pushes logging to a separate process writing to redis or files so process.stderr never blocks. Notes typical tradeoffs: added server CPU (1ms → ~15ms per request) and ability to glue tens of thousands of clients to one machine.
AI-written summary · about
chapters (5)
- 0:00Start
- 3:24Pre-allocate resources to the client for the entire duration
- 4:28When you run out of memory the server gets overwhelmed
- 5:05Don't pre-allocate memory
- 15:47Use file descriptor sharing to do on machine multicast
related talks
23mDecomposition of the Main Thread in Node.js to increase throughput
46mNodeJS Scalibility Tips
53mMastering Node.js Performance
22mScaling Node.js and unblocking the Event Loop
34mNode.js Q&A
49mNode JS
24mWhy Node.js Needs an Application Server
13mConcurrency vs Parallelism in Node.js: Understanding the Event Loop & Worker Threads