Multivendor technology initiatives have a way of gathering momentum simply because they promise to overcome many of the integration and incompatibility issues that turn minor deployment projects into major engineering hassles.
But good intentions are not enough to create an efficient and effective data ecosystem, and no matter how glossy the cover is on the marketing brochures for projects like OpenStack, the fact remains that member firms’ first responsibilities will always remain with their proprietary technologies and the stakeholders that own them.
Now, to be fair, there is a lot to like in the latest happenings of the OpenStack community. At this week’s OpenStack Summit in San Diego, the group clearly showed growing momentum behind the project, with work under way on everything from an integrated computer fabric to virtual machine management and software-defined networking (SDN). At the same time, major backers like Rackspace and NetApp have released additional APIs intended to draw more support for OpenStack-based hardware and software platforms.
And yet, there are some voices who are starting to question whether there is more chaff than wheat here. Forrester analyst James Staten, for one, has pointed out that even though the project is more than two years old, there is only a smattering of actual OpenStack products in the channel. Rival cloud projects like CloudStack and Eucalyptus, he notes, have substantially more options ready for deployment and are said to be already drawing revenue.
Gartner’s Lydia Leong is even more blunt, warning enterprises not to give in to OpenStack hype, but to conduct due diligence on the system just as they would any other cloud management platform (CMP). One of the key drivers for OpenStack, she notes, is vendors’ fear that Amazon Web Services and VMware’s vCloud will come to dominate the cloud market, but that doesn’t mean OpenStack is a better solution from a user perspective. Your overriding goal, she adds, is to deploy a workable cloud API library and/or management tool that provide a unified interface to multiple clouds. And in that vein, OpenStack is still in its infancy.
It is also worth noting that some OpenStack partners support the project more strongly than others. For example, Cisco recently released its own version of the OpenStack software optimized for its Nexus and UCS hardware platforms. To its credit, Cisco maintains that an integrated hardware/software environment is the best way to ensure a fully functional, highly dynamic cloud, but it nonetheless represents a proprietary twist on what is professed to be a fully open cloud development platform. And Cisco is not limiting itself to just OpenStack, mind you. It is also working closely with CA and Citrix to support the latter’s own version of CloudStack.
So clearly, development of the working cloud universe is not proceeding along a linear path, but consists of multiple tangled webs of partnerships, rivalries and ever-advancing technologies. In this way, it is not much different from all the data initiatives that have hit IT in the past.
However, the cloud represents probably the last great chance of a fully federated data universe in which operating environments can truly exist free from the encumbrances of the hardware-driven past. It may be a pipe-dream, but it sure would open up new levels and worker and data productivity.