@Mr_Tom said: In my experience unless a worldwide audience is the target then if its a UK site for UK people and hosted in UK/EU then cloudflare just adds another layer to potentially slow stuff down.
This is an interesting one, I did test this yesterday by removing Cloudflare and reverting back to the provider DNS and from the UK/EU the overall performance based on GTMetrix improved enough to make it worthwhile not using a CDN. As expected not so great for those further afield but they are not the target audience.
I do not wish to derail the thread but one question. If you do not use cloudflare how could you protect you from the more basic ddos? I mean your ip would appear if not using cloudflare and you know etc.
I believe in good luck. Harder that I work ,luckier i get.
@Chievo said: I do not wish to derail the thread but one question. If you do not use cloudflare how could you protect you from the more basic ddos? I mean your ip would appear if not using cloudflare and you know etc.
I guess the question is whether DDoS is going to be an issue for the site. I mean your average site won't need to worry about DDoS attacks. For the majority of site owners, I doubt that DDoS or IP exposure is even a consideration.
@Lee said:
So more of a performance proof of concept I took a large, image-heavy WordPress install, not mine but I had permission and copied it to various locations and then ran it through GTMetrix. Either cPanel/DirectAdmin with Cloudflare.
Of course, there is elements that will affect each result but ultimately two good UK providers vs two good EU providers and performance is better outside of the UK than in. Surprisingly HostMantis vs other is quite poor which I was surprised with. 2.6s to the largest contentful paint vs as low as 1.4s is a major difference that visitors will notice.
Also quite surprised about the HM result. Interesting share, thanks. Regardless, HM has a great bundle of features I grew fond of. On another note HM increased I/O limit again iirc.
@Ympker said: Also quite surprised about the HM result. Interesting share, thanks. Regardless, HM has a great bundle of features I grew fond of. On another note HM increased I/O limit again iirc
Well, this is the thing, it is all very dependent on factors. I mean as mentioned above sometimes Cloudflare gets in the way.
This is me running it again with HM, removing Cloudflare and reverting back to their own DNS.
@mikho said:
No matter how much you investigate, the momwnt you present the results and the clients tests it themself.
The result are always different.
@Lee said:
How much value do you place on location when it comes to hosting websites?
For website such as my blog, I pick location so that my readers experience lower latency.
However, I have worldwide readers so I can't make everyone happy, unless I resort to complicated GeoDNS solutions.
In the past month, I have 38% readers from China, 16% from USA, and 5% from Germany.
My hosting location is Seattle.
While Japan would offer better latency to China without worsening USA much, it's too slow for readers from Europe (20%).
@LeroyJenkins said:
Geo location is important for a few:
Game servers
Geo location matters for push-up servers too.
My January push-ups report, Router Distance section, points out:
Having higher estimated bandwidth allows higher video resolutions.
Most viewers were using a router near them.
Higher bandwidth and resolution can be achieved only if the router is near the viewer.
However, connecting to a nearby router does not imply a high resolution.
Comments
I do not wish to derail the thread but one question. If you do not use cloudflare how could you protect you from the more basic ddos? I mean your ip would appear if not using cloudflare and you know etc.
I believe in good luck. Harder that I work ,luckier i get.
I guess the question is whether DDoS is going to be an issue for the site. I mean your average site won't need to worry about DDoS attacks. For the majority of site owners, I doubt that DDoS or IP exposure is even a consideration.
Also quite surprised about the HM result. Interesting share, thanks. Regardless, HM has a great bundle of features I grew fond of. On another note HM increased I/O limit again iirc.
Ympker's VPN LTD Comparison, Uptime.is, Ympker's GitHub.
Well, this is the thing, it is all very dependent on factors. I mean as mentioned above sometimes Cloudflare gets in the way.
This is me running it again with HM, removing Cloudflare and reverting back to their own DNS.
TTFB (58ms) - FCP (0.7) - LCP (1.5)
And with Cloudflare back on.
TTFB (0.6s) - FCP (1.1) - LCP (1.5)
All very interesting, I am spending way to much time on this
Haha, fair enough. Thanks for sharing, mate
One can really spend lots of time on this :P
Ympker's VPN LTD Comparison, Uptime.is, Ympker's GitHub.
No matter how much you investigate, the momwnt you present the results and the clients tests it themself.
The result are always different.
“Technology is best when it brings people together.” – Matt Mullenweg
Words of wisdom
Ympker's VPN LTD Comparison, Uptime.is, Ympker's GitHub.
It is always variable, never going to change that.
If anything it does prove that a CDN can in fact add more time to the page load. But again, depends, variables, etc..
Learnt a lot in the last 48 hours, that's the real benefit, for me at least
How is your friend's Blocksy site getting along, anyway? Performance is fine?
Ympker's VPN LTD Comparison, Uptime.is, Ympker's GitHub.
How is your friend's Blocksy site getting along, anyway? Performance is fine?
Ympker's VPN LTD Comparison, Uptime.is, Ympker's GitHub.
Guess which site I have been moving around and testing
But yes, she is good with it, loves Blocksy. I have started to use it on a couple of sites as well.
This whole discussion started when someone suggested her site was slow.
Ah, I see
Does not seem "slow" to me, though. Clearly they haven't seen unoptimized Divi Installs yet :P
Anyway, good luck optimizing!
Ympker's VPN LTD Comparison, Uptime.is, Ympker's GitHub.
Geo location is important for a few:
Anything else - 0 care for location.
It's not, but I then decided to demonstrate and two days later still thinking about it
For website such as my blog, I pick location so that my readers experience lower latency.
However, I have worldwide readers so I can't make everyone happy, unless I resort to complicated GeoDNS solutions.
In the past month, I have 38% readers from China, 16% from USA, and 5% from Germany.
My hosting location is Seattle.
While Japan would offer better latency to China without worsening USA much, it's too slow for readers from Europe (20%).
Geo location matters for push-up servers too.
My January push-ups report, Router Distance section, points out:
Accepting submissions for IPv6 less than /64 Hall of Incompetence.