Skip to content

Comment on Scaling node.js to 100k concurrent connections

Comments

Looks like author is not aware of some cuncurrency problems, deadlocks etc. Backend/database might not scale to 100k concurrent connections so easily.

This is where NodeJS really starts to shine - persistant connections and background operations let you do a whole bunch of cool stuff to mitigate that.

In my case I have entire db tables and collections replicated in memory and kept in sync via redis pubsub, and the 100,000s of concurrent users I have are all sharing just a few dozen persistant redis and mongodb connections between them.

Scaling the backend is a lot easier than dealing with concurrent front-end connections!

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.