Latency Test Details

🖖

Hello fellow nerd *tips fedora*🕴

So ya want to find out how our latency tests are done in technical detail? Well you have come to the right place.

I'll start by saying our Latency Tests are not perfect, but they were as good as what I could think of to display the users current latency to servers worldwide and provide a prediction on what their latency would be if they decided to join our network.

So how do they work? When a user visits our website we have Javascript that will run client-side and setup WebRTC to servers we host worldwide. Once this WebRTC session has been established the client will send/receive UDP data to the servers worldwide and measure the latency. For the results in the Latency Test table the lowest latency result for each server is displayed. This is how the users current ISP latency is displayed.

Now how do we predict what the users latency will be if they decided to join our network? We are able to do this by measuring the clients latency to perth.ozot.com.au (In our NextDC P2 Data Centre). This gives us a decent prediction of what the clients latency will be to our Perth Data Centre (not perfect). Once we have this figure we are able to + this latency figure to our server-side WebRTC latency results that run to the exact same servers worldwide. Using this data we are able to predict the users latency on our network by adding their latency to our Perth Data Centre to these figures. It isn't perfect, but it is an ok prediction for what the user should expect.