James Water has an interesting take on the Rackspace announcement. My comments follow:
My Take: Its early enough to grow their own ecosystem, but it will require sincere focus and value-ad to justify the differentiation
Rackspace’s API is different for a reason. There are certain features (like static/unique IPs) they wanted to support that EC2 did not.
All the same that puts the onus on them to continue to create value ad differentiation for their cloud until there are numerous ISV’s building natively on the Rackcloud instead of porting over to it—that is until they have a unique market footprint, and their own cluster of long tail users who prefer their cloud for those architectural features. This is a partial departure from their service and quality branding—its compute quality and feature centric.
I agree with James in that is too early to declare winners and losers. But I'd say that the window is closing very fast. You got VMware, Amazon, Microsoft in the leading spots for 1-2-3. The fight now is to be #4 or to unseat one of the incumbents.
The issue of compatibility vs innovation is one of the real difficult ones any platform provider faces. What I see is this: why not just embrace the EC2 API's? Launch a server is launch a server. You can always extend to support differentiated value. This would get you a whole bunch of ISV's for free, then let the customer drive extensions by beating up the ISV's.
Embrace and extend. That's a strategy you don't see working often. Oh wait.