According to Chrome the difference is "waiting" for the server response (47 ms Vs. 473 ms). You can tell that nginx is configured very differently between the two response types just by looking at the response headers.<p>I suspect the author set out to prove this and then made the environment fit their agenda. Nothing to see here. I'm sure Spdy is faster than HTTP/1.1 but nowhere near what this site claims.
>For best results, disable plug-ins<p>Why? It's perfectly reasonable to be more concerned with real life behaviour than with benchmarks under optimal conditions.
I don't want to be a killjoy but here I got HTTP faster than HTTPS<p>3 tests<p><pre><code> 63%
59%
60%
Mac OS X 10.9.5
Google Chrome 39.0.2171.71 (64-bit)
Google Chrome Helper disabled</code></pre>
Sorry, everyone. I'm the site's author, and I increased the ulimit and worker_connections to 4096, so you should see better perfomance and no more errors now. My apologies...I'm a web guy, not a networking guy. Rookie mistake.<p>And yes, you did break it :/
Of course SPDY keeps the same TCP connections alive, the encryption and even TLS handshake time is insignificant compared to the latency in creating a new connection for each HTTP request.
Looks pointless to me. Because choosing between turning your resource to user with http or https is not about speed. There are other considerations on the first place.