Roy Cooper roycrxtw

This set of benchmarks explores the impact on a Node.js HTTP API server when some responses take more cpu than others. The requests are to a JSON page that calculates a random fibonacci sequence between 1 and 30. The server is using node-perfectapi with a feature that scales the http server across available cpus, while leaving the user's API code in the main Node.js process.

In the red series, the http server is running in the same process as the Fibonacci calculations. In the green series, the handling of all of the http web server stuff has been moved to separate node workers. In both cases, the Fibonacci calculations occur in a single thread in the main (master) Node process.

Now obviously the red series is going to be slower than the green series (because it is restricted to a single cpu for both http requests/responses and the calculati

What is your folder-structure preference for a large-scale Node.js project?

0: Starting from Rails

This is the reference point. All the other options are based off this.

|-- app
|   |-- controllers
|   |   |-- admin

	before(function (done) {
	/**
	* Mocha is extremely zealous about trapping errors, and runs each test
	* in a try / catch block. To get the exception to propagate out to the
	* domain's uncaughtException handler, we need to put the test in an
	* asynchronous context and break out of the mocha jail.
	*/
	process.nextTick(function () {
	// disable mocha's error handler
	mochaHandler = process.listeners('uncaughtException').pop();

	// A very basic web server in node.js
	// Stolen from: Node.js for Front-End Developers by Garann Means (p. 9-10)

	var port = 8000;
	var serverUrl = "127.0.0.1";

	var http = require("http");
	var path = require("path");
	var fs = require("fs");
	var checkMimeType = true;