When comparing Cloudflare and Incapsula, I choose one tool to use for intermittent testing. Google Developer PageSpeed Insights. I would look at other cumulative statistics like Google Analytics, server bandwidth, Webalizer and AWStats and server outage resilience (would the cached website still be up when the server was down?). Overall ease of set up and use is also a consideration. Each CDN provider comes with it’s own set of tools, measurements and analytics, though these aren’t thoroughly compared because I was really just interested in which service would lighten the load on my servers the most and increase page speed.
See how the websites did by checking out the Website Profiles on “Content Delivery Networks – The Need for Speed” over at Champlain Websites.
This was true of both services, which was kind of a bummer. I had two of the websites I used for testing set up to go without the www. I had to make the switch and leave plenty of time for propagation before I could begin testing. If my theory of server outage resiliency works out how I think it will, I might never go wwwless again.
CDNs can be Complicated
A Records, MX Records & CNAMEs. I didn’t want to deal with this… With Incapsula, I had to. It was in the instructions. In my case it was at the hosting level, not the registrar. Cloudflare was easy. Change the NS at the registrar level and you’re done. Though I have yet to test that with email.
This is where CDNs become sticky. Email addresses from a domain need to be considered heavily when making changes to your website. I always recommend separating your email from your website as much as practical for this reason. GMail is my goto, but that’s a different story… Incupsula’s instructions for CNAME and A RECORD changes are easy to follow for someone who’s familiar with DNS Zone settings. For others, taking the time to carefully read the instructions and patience are required. Patience in general is required when making changes to these records since the TTL (Time To Live) time is usually set for 4 hours on these records (14400 TTL = 4 Hours). You can edit the TTL at the same time you make the changes to the record and set it to 3600 (1 Hour), so that if you need to change it again, it’ll only take an hour to propagate.
Can I Reduce Website Bandwidth?
The results on this are mixed so far. Testing is ongoing, and at the time of writing this, my biggest bandwidth hog is one month into testing with no positive results. Of the four original websites I tested, one on Cloudflare and one on Incapsula saw big reductions in bandwidth usage. All five were switched to the opposite CDN to see if bandwidth changes on any of them.
Slightly less results with Page Speed, comparable results with bandwidth.
Ease of use makes it the winner.
Excellent product, but the implementation is too much of a pain. Documentation is pretty good. Support is fast and good. Still worth looking at it for other reasons.