Servers: Want some BENCHMARKS? (16)
So, I've been noticing that there's a lot of interest in which servers, tracking systems and so on are the fastest - but as far as I know, there haven't been a lot of comparative benchmarks done to find out that information for sure.
Is that something which you guys would be interested in seeing?
I'm thinking tests like:
- CPVLab vs Prosper vs POFpro: redirect speeds and server load.
- BeyondHosting vs Storm on Demand vs Linode vs Amazon EC2 (for laughs) - performance under real-world affiliate-style load
So, does that sound like A Thing That Would Be Helpful? Let me know, and if it does, I'll get started on organising it.
This is something I've always wanted to do but never had the time.
Would be some very interesting info.
If anyone wants to beat on a BH machine please start a support ticket so we can assist with the benchmarks!
Tyler - I was planning on getting in touch about just that - will do 
That is a really good idea and thank you!
But please if you measure redirect speeds, do a real comparison. I can not see these "under 100ms redirects" comparisons anymore... In the real world it doesn't matter if it's 100ms or 500ms, more important is that your landing pages are highly optimized for loading speed. And then a difference of half a second is nothing...
But the rest will surely be interesting 
Don't worry, the tests will be rigorous!
Although you might be surprised how much half a second affects your CTR - in my tests, saving half a second increased CTR by 50%...
Yep, but you tested the page loading speed, that of course has a huge impact 
I was talking about these tests where people measure pure redirect speed, i.e just the time the server needs to return to the client and have dns resolve. Like Mobafftracker says they have redirect speeds less than 100ms, but that doesn't do a thing if your lander is loading slow.
Does having the tracking on a server other than the pages served effect speed?
Does using web fonts help?
Does using CDN for the images and or vids help?
Are Meghan Fox's tits real?
CDNs are something I'm interested in - I have used landers with one main background image and when I offloaded it to a CDN the page load times were ALWAYS slower. The extra time taken to request the image from the CDN seemed to make it redundant. I'm not sure why it was happening like that (using Amazon).
Originally Posted by zeno
Also the first time you load the page (if you use pull it was I think) takes longer because amazon downloads the image, copies it to the regional CDN server that is closest to you and then serves it.
Yeah the initial download made things slower, but it was always slower on repeated tests too. In the end I just cbf dealing with it as I hate the Amazon system, it's so clunky to navigate through when all you want to do is put a damn image on a CDN. Buckets, distributions, upload, change permissions, make public, get links, blah blah. Why no drag -> drop -> given link.
Originally Posted by zeno
Great idea, but to make the tests fair should not have a host or their staff helping with the benchmarks.
CDNs - right, good call, everyone. I know there's a lot of interest and a lot of mixed results people are getting from CDNs - I'll put them right at the top of the pile to test.
@sandyone - all good ideas! If you can set up a test environment for the last one, I'm happy to give that some *rigorous* examination...
@nissangtr - I agree, it should either be "all" or "none". Although there is an argument for saying "well, if one host will help me optimise my speeds and another one can't be arsed..."
Were these results in ?
Home > Programming, Servers & Scripts > Hosting, Servers & Security