This is a very comprehensive post about the tremendously important topic of lock-in with respect to IaaS and PaaS. Well done. Please, I'd like to respectfully add to the discussion.
In the long-term, I see traditional web app developers/builders gravitating toward open source PaaS solutions rather then making direct use of IaaS solutions, for several reasons. Keeping to the context of your post, PaaS is a winner because it is a layer of abstraction that, if done correctly, frees apps from things that might bind them to an IaaS provider. An emerging example of this is EngineYard's PaaS offerings. Today, you can deploy the same Ruby on Rails app to Engine Yard's AppCloud and xCloud platforms, PaaS offerings built upon two very different IaaS layers (Amazon and Terremark, respectively). Who's to say EngineYard wont announce partnerships with other IaaS/hosting providers.
Am I suggesting that IaaS is going to lose to PaaS? Quite the contrary. What I envision is that PaaS and IaaS providers become more and more interlinked with and dependent on one another. PaaS providers build upon IaaS offerings, thus aggregating significant revenue for the IaaS providers. When one IaaS layer is better in terms of scalability, price, etc., the PaaS provider focuses on that IaaS until the others catch up. In turn, the symbiotic relationships among PaaS and IaaS vendors inevitably drive cloud standards that more easily and cost-efficiently support their coexistence. The end result is cloud app platforms in which the only lock-in layer is the programming language of choice (Java, Ruby, .NET, etc.).
I hope my thoughts add something.
Steve Bobrowski
The Cloud View ( http://www.thecloudview.com)
The best of ZDNet, delivered
ZDNet Newsletters
Get the best of ZDNet delivered straight to your inbox
