As you may know, we’re pretty big fans of NodeJS here at Raygun – we run it in production and love talking about it! With the recent release of NodeJS 6.0, we decided to take a nosey and see how it performs compared to older versions of Node.
As before, we’ve followed the same pattern – Ubuntu VM running AB pushing a large number of requests (100 concurrent connections) into the Raygun API to see how many requests per second it copes with – usual disclaimers apply – these numbers relate only to this one blog post, they don’t aim to be a complete overview of how well NodeJS performs.
This test only stress tests a very small portion of what can be tested.
Let’s look at some NodeJS performance numbers!
“Huh-what” you’re probably thinking, is this the opposite to what is meant to happen?”
You’re quite right – but, and this is perhaps the most important part, this code was built against 0.10.* and has been optimised for it.
What happens then if we compare a basic Express app that I just created?
The slight slowdown that has been witnessed in Node over the last few versions is gone – we’re back to where we were, in terms of Express speed anyway.
NodeJS 6.0.0’s huge focus was on speed and it shows – it’s jumped up way past 5.0 in my basic testing – always nice to see an improvements that can mean real world savings in hardware costs!