> yikes those numbers are terrible something is really wrong there.
> secondly responses will be much slower if you dont specify
> the Content-Length, (use res.send('hello world')), but yikes yeah
> something is very wrong with that, normally you would have to
> explicitly try to make express that slow, so either the network hops
> are brutal or something else is wrong

> On Nov 2, 1:42 pm, Gustavo Machado <machad...@gmail.com> wrote:
> > Hi,

> > I would like to start a series of benchmarks of a few apps we have
> running
> > with node.js so I decided to run a baseline benchmark on a vanilla
> > express.js application with the following route:

> > app.get('/', function ( req, res ) {
> >   res.end("hello world");

> > });

> > I ran ApacheBench like this:
> > .\ab.exe -n 100http://one.org:3000/
> > .\ab.exe -n 100 -c 10http://one.org:3000/
> > .\ab.exe -n 100 -c 100http://one.org:3000/

> > (one.org is mapped in my hosts file to an EC2 Large instance)

> > I repeated the tests several times, and results were pretty consistent:
> https://gist.github.com/4004085

> > So my questions are, do these numbers make sense? And is this suggesting
> > that in the last test one request was waiting for 24 seconds? Why?

> > Is there a better way to benchmark an express.js app?

> > Thanks,
> > Gustavo