Is traffic shaping working on the Nexus S tests?

Tested the same site on both the Nexus S and the iPhone but get very different load times Nexus - ~2s, iPhone - ~7s

Now I know the iPhone uses a UIWebView rather than the browser but I don’t think that explains the differences.

Nexus: http://www.webpagetest.org/result/121228_DX_d27d025b854f58dda915a71150e3d4a7/

iPhone: http://www.webpagetest.org/result/121228_JZ_c100646a583ed1d3789db3d4e37629c5/

Any ideas?

Argh, good catch, thanks.

I have a machine running dummynet between the phone’s wifi access point and the router and it was incorrectly configured to shape a /24 block instead of /16 and the iPhones and Android phones are on different subnets.

It should be fixed now - sorry about that - http://www.webpagetest.org/result/130102_JZ_cdf5e2e5acd8596fdd0e9098fa4774dd/

It’s pretty sad that I’m nostalgic about the relative ease of testing IE on Windows (given how much of a black box IE is) :frowning:

No worries… Got enough data from the iPhone agent to write my blog post, but will retest using the Nexus agent too…

Yeh, there’s got to be a better way of testing all this stuff!

Have you seen this for generating HARs via Chrome Remote Debugging - https://github.com/cyrus-and/chrome-har-capturer