MyBB Internal: One or more warnings occured. Please contact your administrator for assistance.
WebPagetest Forums - Which test node has the lowest latency (RTT)

WebPagetest Forums

Full Version: Which test node has the lowest latency (RTT)
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Hi,

i've been testing all afternoon, and the results ive got, are very interesting.

However what i have found is that i'm not getting 28ms RTT, its actually nearer 50ms, and that is obviously an artifact of the test node (and its connection to the internet/ISP).

So i'd like to use the node with the lowest inherent RTT.

Which one has that (without me testing all of them to determine it).

Would it be your Fios goodness Pat?

@hsiboy
The EC2 nodes tend to have very good latency (Oregon, Ireland, Singapore and Sydney) though you have to watch out for the isp->CDN latency too. Dulles tends to work really well as well but it does have ~4ms of latency on the connection itself (should be subtracted out from the traffic shaping).

What are you measuring RTT to? Your server or a CDN edge node? Some of it will depend on the natural network latency from the physical location to the server being hit (and in the case of a CDN it will also depend on how good they are at picking the best edge node).

You can try running at "Native" speed which will show you the natural latency on a given test node to the servers being tested.
Hi Pat,

i was using the Dublin node, but found i couldn't get better than about 50ms latency in my tests (using native). So despite it being set to DSL 28ms it couldn't get that low.

I'll repeat the tests from your basement ;-)

(07-23-2013 11:04 PM)pmeenan Wrote: [ -> ]The EC2 nodes tend to have very good latency (Oregon, Ireland, Singapore and Sydney) though you have to watch out for the isp->CDN latency too. Dulles tends to work really well as well but it does have ~4ms of latency on the connection itself (should be subtracted out from the traffic shaping).

What are you measuring RTT to? Your server or a CDN edge node? Some of it will depend on the natural network latency from the physical location to the server being hit (and in the case of a CDN it will also depend on how good they are at picking the best edge node).

You can try running at "Native" speed which will show you the natural latency on a given test node to the servers being tested.
Latency to where though? Was it another server in Dublin somewhere? Otherwise you're going to have natural latencies in the network itself.
Reference URL's