Why Java should dominate the cloud-scape
Sun has all of the assets necessary to dominate the cloud-scape.
Underlying all the cloud possibilities is a reliance on more traditional application deployments. So far virtualization and VM images are the dominant factor in both packaging and consumption.
This leads to an interesting quandary for those who want to embrace the cloud full-on, requiring a developers to adapt to the new infrastructure design.
Considering the challenges of clustering in general and then consider the challenge of clustering from the enterprise to the cloud in a federated manner. VMs have become much more manageable and easier to use, but there are still complexities. Many monitoring tools don't function well with VM systems. Or consider the potential annoyances associated with backup and recovery. Easy in theory, but in practice you need to have your ducks in a row.
The right answer is to move to a cloud infrastructure that runs in a virtualized layer, completely invisible to the end-user. And it should be on the user's operating system and development platform of choice. This is an unlikely scenario for Windows, but highly likely for both Linux and Solaris.
As the enterprise platform of choice, Java seems like the logical option to underlay cloud infrastructure. To the extent that Sun is interested I would say it's a decent bet, but efforts to date have been middling at best. And there continue to be some mixed messages-including some changes to JEE 6, which introduces some features that seem not very meaningful for cloud-based deployments.
Sun has an amazing combination of software, hardware and storage and so I would I would expect a more cohesive story. In fact, I would argue that of all the BigCo vendors, Sun has the best chance of becoming a meaningful cloud vendor.
At the moment Java is really only available if you are running in VMs. There is no cloud infrastructure that lets you natively deploy and run Java applications as you would deploy Python on App Engine or file downloads on S3. If we assume that .NET is the other big development platform (let's take PHP out just for the moment) then once again there is no direct deployment strategy. And Microsoft remains fairly clueless despite some recent news.
The big question is whether Sun can take the dominant position vs. IBM, HP, and Red Hat.