I have just decided to do a little experiment with this site, because some of my clients' sites are running slowly. I wanted to check all the angles.
Firstly, I have been looking at template bloat, javascript bloat, css bloat, image bloat, WYSIWYG editor bloat...oh, you get the idea.
Also caching - native Joomla and add-ons, e.g. MySQL caching. Also the drag added by SEF URLs and plugins. And various other flora and fauna. It all adds up.
I posted about the editors last week, but this post is about my hosting speed test.
I made 2 copies of the this site (since I have custodianship of it). Then I popped the two copies on two different U.S-based dedicated virtual servers (DV1 and DV2 below). This site itself is hosted at Rochen, whose servers are in America.
According to tools.Pingdom.com (site timer), the home page load times were:
1. This site: load times varied from 1.95s to 15.98s
2. DV1: load times varied from 6.45s to 10.6s
3. DV2: load times varied from 4.63s to 8.56s
This is measured from Pingdoms servers in Dallas, Texas and all the hosting servers are in USA.
Measuring the same pages from my trusty browser in Sydney, using Firebug, I got the following load times (over two attempts)
1. 20 seconds (second attempt 12 seconds)
2. 27 seconds (second attempt 18 seconds)
3. 33 seconds (second attempt 22 seconds)
So I think we can see the influence of Australian comms (and my antivirus program!)
Anyway, on doing a traceroute, it turns out that DV2, although faster from US to US is slower from Aus to US because it's on the East Coast and the pings goes a long and (to use a technical term) wiggly route to get there. DV1 is West Coast server and Rochen is Dallas, and the route is much more direct.
So it makes you think, doesn't it, about where your servers are?
Anyone got comments? Am I all hot air?