Have you ever asked yourself what your website or application’s network latency is to various regions and cities around the world? If you have, this blog post is meant just for you.
First off let’s define what Network latency is:
Network latency is the amount of time packets take to travel from a source to a destination on the internet. Network latency is also referred to as rtt (round trip time) or round trip latency.
By extension, therefore AWS latency refers to the time it takes for data packets to propagate between a specific AWS region and a client. Whenever a client or an end user accesses a website or an online application hosted in an AWS region, the time that it takes for their browser to communicate with the AWS cloud instance and receive the first byte counts as AWS latency.
In this blog post we are going to take a look at the network latency map from Datapath.io to see the latency from different AWS regions to globally distributed clients. The map can also be used to see the network latency from other cloud and bare metal service providers through various downstream transit providers to end users, that however will be covered in an upcoming blog post. Here we are only going to look at the network latency from AWS Virginia (US East-1) over the default AWS route to end users in different parts of the world.
The screen shot below shows the network latency from AWS Virginia (US East-1) through the default AWS route. The region is marked out by the small bluish-green dot to the right of the picture.
A word about the navigation panels at the top: In the top right corner, you can select the cloud or bare metal service provider you want to check network latency for. Right next to it you can choose different AWS regions or downstream transit providers. You can also search for individual networks in the search bar directly below it.
Okay so back to the latency map: The US shows greens all over the board (hint: green is good). However, you can change your latency tolerance from the slider at the top right. Right now it is set at 100ms which is tolerable for most service providers.
However, go a little to the right and we start getting reds all over the place. Average AWS latency to the Russian coastline from US East-1 is 229 ms, 258 ms to the Chinese coastline and 282 ms to the Chinese mainland.
You can always drill down to see the individual AWS latency for specific regions and cities. Case in point: Beijing which registers at 302 milliseconds from AWS Virginia.
Africa has a mix of red and brown Network latency indicators. The highest network latency in Africa is to Madagascar at 264ms followed by the southern tip at 235ms on average. India and Iran also suffer from high network latency of 251 ms and 191 ms respectively.
The United Kingdom and western EU are within the 100 ms latency mark. But as we move to the eastern EU and the middle eastern region the latency indicator starts shifting from brown to red.
South America registers 90 to 145 ms on the average AWS latency scale from US East-1.
Australia, out in the middle of nowhere, is all red with an average AWS latency of 227ms and 264 ms on both coasts.
So what does it all mean?
Network latency is a pretty good indicator of page load times or the speed of your online application. The quicker your website or app loads and responds to customer actions the more engaged they are, the more engaged they are the more likely they are to take a conversion action.
AWS latency from different AWS regions to geographically distributed clients or end users is also a pretty good indicator of page load times and the speed of online applications. Higher AWS latency translates into higher page load times and consequently slower online applications. AWS latency to different regions is crucial for service providers with VPCs hosted in just one AWS region. If we take the example above, end users in Asia or Africa of an online application hosted with AWS Virginia US East-1 will experience higher Latency and slower applications. Geographically distributing instances of a service over multiple AWS regions can mitigate the effects of high network latency to some extent but that can also result in a nice big bump in costs.