On May 21, 2015, at 09:09, Ian Schenck <ian.schenck@gmail.com> wrote:

Gelin,

I'm loathsome to share my data as benchmark data.

Ian,

Thank you for sharing, thank you for doing a realistic test, and thank you for qualifying your results as heavily as they really should be :). I only wish that more publishers of hype-fest microbenchmarks would be as scrupulous and methodical.

With that disclaimer, I've attached a graph of mean and standard deviation in our particular test.

<Request-Response Latency.png>
Run with:
ab -c 1 -n 100
ab -c 10 -n 1000
ab -c 25 -n 5000

I don't know if I have complete output from ab anymore.

One thing I always feel compelled to point out - 'ab' is a terrible benchmarking tool and the numbers it outputs are, often as not, just completely wrong. It really is "apache"-bench, meant to work for validating apache configurations, not a general HTTP testing tool. I have personally had much better success with the very simple "httperf", although there are other tools that are probably better than that.