When designing an solution involving any cloud vendor, you need to be aware of the network latency between your users and the cloud. For Azure, this means putting your objects (VMs, web sites, storage, etc) in the data canter closes to you. But which one is that?
I just came upon a neat page that can help: http://linkis.com/azurewebsites.net/Jaybw. This page plots a nice looking graph of the latency between the client and the 15 existing Azure data centres around the world. After a few tests, my graph looks like this:
There's also a nice table that goes along with the graph:
As you can see, the latency between my office and Azure favours the Western Europe data centre in Amsterdam. I had expected Dublin (North Europe) to be faster or at least very close – but was in fact slower. . Not surprisingly, Brazil, Japan and Australia are a lot further away. I was also surprised that the times to South East Asia and East Asia were faster than both West US and East US.
What do your tests show?