As a follow up to the previous post on installing CloudFlare, the implementation was completed on our corporate site with some astonishing results. I must admit, I was skeptical, I was not sure it would really deliver the promise offered. Essentially, there are a lot of features they add for free, and you really can not help but wonder what is thier business model, how do they profit and what do they need to do. I was expecting a range of endless offers for upgrades, new features and such – all at a cost but nothing.
So the end result, the real reason for doing this was an improvement on the speed. We are using webpagetest.org and have been doing so since before the Cloudflare CDN implementation. So while there might be some reservations by others about the accuracy of that platform, and I suspect any platform will have its detractors and non-believers.
Table: Results following the CloudFlare CDN with full implementation of CloudFlare
Table: Results before implementing the CDN CloudFlare
The comparison between the two is almost not to be believed, The speeds we were getting before the implementation were terrible, there were complaints from everyone, users internally and externally as well. Some of the settings have not been fully enabled, for instance, we chose not to have the asynchonised loading of javascript, which there is a lot of embedded in to the homepage.
Overall, the CloudFlare CDN has saved 254 GB of bandwidth that our servers did not have to produce. In turn, our processors also did not have to work for. The suprise easter egg is the number of threats that the CloudFlare CDN has also blocked, what was also a surprise was that the threats were originating from Australian IP addresses. A full table of the results is included below.
