Much has been discussed on Hybrid Clouds. One way to think about the problem of what should be in Public vs. Private cloud is what would Amazon do moving from Public Shipping companies to add capacity with a private delivery fleet. Why? Because sometimes thinking how to solve a problem from a different perspective can show you a different way to think about the problem.
If you were going to move some of your Cloud Compute out of AWS what should you move? The easiest or the most valuable? Another way to think is what could you move to your Private Cloud that would give you the most savings, leaving the rest to run in AWS.
Let me use local shipping as an example. Here in Seattle UPS, Fedex and other local contract companies deliver packages for Amazon. If Amazon was going to add private delivery to the mix what should they add? Those thing that are the easiest? Those things that are the most valuable? You could, but that doesn’t necessarily get the maximum value out of a local private delivery service. If Amazon has 10 trucks for routes and knows the shipments coming through, why not pick those shipments that can have the most savings vs. shipping with the Public carriers? The complexity is a good problem for compute algorithms to calculate those shipments that can be handled in a day with the at the lowest cost and the most amount of freight saved. This difference is what ends up in Amazon’s pocket at the end of the day, and provides future negotiating leverage with the Public Carriers.
Sometimes what people are thinking goes into a public cloud vs. private is not a clear financial reason. If you cannot run it cheaper in house, then move it back to Public, and put something else in the private cloud.
Just something that hit me after reading the Everything Store book.