Manchester so much to answer for

Morning,

I’m hoping someone can help me.

I’ve been running a few tests and am getting some weird results. Can anyone explain the dramatic differences and rational behind long load times?

https://www.standardlifeworkplace.co.uk/workplace
From: Manchester, UK - Chrome - Cable - First view load time 15.267s

https://www.standardlifeworkplace.co.uk
From: Manchester, UK - Chrome - Cable - First view load time 56.871s

Test ran with ‘Ignore SSL Certificate Errors checked’

https://www.standardlifeworkplace.co.uk/workplace
From: Manchester, UK - Chrome - Cable - First view load time 4.884s

https://www.standardlifeworkplace.co.uk
From: Manchester, UK - Chrome - Cable - First view load time 4.699s

Any help/thoughts appreciated, thanks

I always test with the cable connection in Dulles VA because over the years I have found it to be the most consistent internet connection in WPT. I also only test with IE 11 for “worst case” scenario.

Here’s a typical test on you, I ran 4 or 5 getting about the same results:

http://www.webpagetest.org/result/160114_9N_11EX/

Thanks Anton,

Agree in principle re: your “worst case” scenario and IE11.

Unfortunately running the test from Dulles VA still shows the site taking 13.886s on Chrome. So in this case Chrome is the “worst case”!!

http://www.webpagetest.org/result/160115_7N_HQ8/

I should be clear I’m not saying that this is the fault of where I’m running the test from. I’m more leaning toward some issue with HTTPS but I’m not sure what.

Thanks

::::: ::::: ::::: ::::: :::::

UPDATE…

So I got a little obsessive after my initial post above and decided to take a little world tour. It might help if I mention I’m based in Edinburgh, Scotland UK to put this into some sort of perspective.

The following we’re all ran on 5x, Chrome with a Cable connection. Remembering I’m based in the UK its surprising Ireland comes out worst. I should also make it clear that in reality on viewing the page load in Chrome it takes no where near these times to load.

Ireland - EC2 61.331s
Shanghai, China 48.645s
Australia - EC2 46.141s
Amsterdam, NL - Go Daddy 27.593s
Moscow, Russia 21.163s
Johannesburg, South Africa 16.785s
Ashburn, VA USA - Yottaa 14.927s
Dulles, VA 13.886s
Frankfurt, DE 13.629s
San Francisco, CA USA - Yottaa 12.174s
Sydney, Australia (Servers Australia) 11.681s
Brazil - EC2 9.2s
Hong Kong, China 8.869s
Petah-Tiqwa, Israel 6.938s
Montreal, Canada - Akamai 4.969s
New York, NY USA 3.531s
Brussels, BE 3.517s
Paris, FR 3.418s
Prague, Czech Republic 3.104s

Are you using EV SSL Certificates? The long gaps in Chrome are the certificate verification (which Chrome only does for EV certs).

They only show up in the waterfalls for Firefox right now: http://www.webpagetest.org/result/160120_XR_ca2f64c77a0c6742a34ac37659c98237/

Requests 1,2,4,19 and 30 are all certificate checks in this waterfall: http://www.webpagetest.org/result/160120_XR_ca2f64c77a0c6742a34ac37659c98237/1/details/

Depending on the performance of the OCSP responder it can cause huge gaps like you see from time to time. You won’t generally see those in your local testing since the certs were recently validated and it isn’t checking them every time.