Skip to content

Instantly share code, notes, and snippets.

@adohe-zz
Last active August 29, 2015 13:56
Show Gist options
  • Select an option

  • Save adohe-zz/9105755 to your computer and use it in GitHub Desktop.

Select an option

Save adohe-zz/9105755 to your computer and use it in GitHub Desktop.
concurrent connections vs connections per second
concurrent connections and connection per second are two different beasts.
To handle 1m concurrent connections a traditional synchronous server would
create 1m threads (or worst, processes) with all the stacks and memory needed.
Node handles that in a sngle thread, but the more connections there are, the
bigger latency per request is need to handle them. because it's single-threaded,
you can however utilize the cluster module to use more than one cpu core on your
machine.
But to handle 1m connections per second, this is about latency, and this you
can only get for the price of higher resource usage. cluster module, horizontal
scaling over several physical machines. etc. node is not different here form
other tools. In fact typicaly it has same latency pattern as rails or php-based
solutions, if you use them the same way. Node's strenghth is the ability to handle
concurrent connections with much lower footprint of memory.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment