There are good technical reasons for this. Anything latency or throughput sensitive is better done within the same datacenter. There have been submissions about

There are good technical reasons for this. Anything latency or throughput sensitive is better done within the same datacenter. There have been submissions about an ffmpeg as a service company and a GPU over TCP company on HN recently that would significantly benefit from 'same cloud'.