While most of the technology world was watching Steve Jobs launch the iPad, some (unfortunate) others were listening to Oracle executives lay out the roadmap for the newly approved Oracle-Sun Microsystems behemoth. Most of what they learned was what they already knew, thanks to months of Larry Ellison leaking some plans and presenting others to the European Commission: Oracle will sell integrated hardware-and-software systems, will invest more in MySQL, and will keep most of Sun’s product lines (albeit as a sort of younger brother in many cases).
It wasn’t until the following day (Thursday) that we finally learned Oracle will not be pursuing Sun’s once-much-anticipated public cloud platform. This decision might prove shortsighted on Oracle’s behalf, and it certainly is a setback to cloud openness and interoperability. While others are taking baby steps toward these goals via open APIs and the like, Sun designed its entire cloud around these principles.
While Oracle has positioned itself well for major sales in the short run, the company could be out of luck if the cloud shift happens to the degree many think it will. The company professes a desire to be the IBM of the 1970s, while IBM has its eyes set (at least in part) on the future, with a variety of cloud services, including a rumored IaaS offering. Additionally, Microsoft has Azure (and a close alliance with HP), while the Cisco-VMware-EMC triumvirate seems to have IaaS ambitions, as well. If IaaS becomes profitable on a large scale, Oracle will be the only systems vendor, or collection of vendors, without a cloud play.
By ditching Sun’s volume server business, it’s arguable that Oracle isn’t positioned to sell to cloud-service providers, either – unlike HP, IBM, Dell and Cisco (including with its new NetApp partnership), all of whom target service providers and organizations with web-scale operations. Oracle’s cloud-computing pony appears hitched to the private cloud, which some believe is here to stay, and which others think is too expensive a proposition to catch on.
But it’s not just Oracle that could hurt as a result of its decision to kill the Sun Cloud. Anyone truly concerned about openness and interoperability also might suffer, as Sun designed its cloud around a whole collection of open components that were supposed to let it tie into into internal systems and, ideally, other clouds. The best cloud users are looking at now might be Cloudkick’s newly announced monitoring and analytics service, which tracks a variety of performance metrics across seven cloud services, or RightScale’s flagship offering, which lets customers manage and launch virtual servers on multiple public clouds via a single dashboard. VMware also is working hard – and successfully – toward its vision of hybrid clouds that can leverage a network of VMware-based public clouds. But choosing from among a pre-defined collection of clouds is not true interoperability of the type toward which Sun appeared to be striving.
There’s no guarantee that Sun could have carried through on its vision, that others would have followed by using its open source components or that it would have been financially successful, but now we’ll never know for sure.