Are you interested in learning about making OpenStack VDI and DaaS a reality? Watch our webinar for all the details.
Of all the well-known names in virtual desktops, two consistently take the cake: VMware and Citrix. Other big players like Microsoft, Oracle, and Red Hat have been scratching at their heels for years, but all have yet to make a significant dent in the VDI duopoly that is defined by VMware and Citrix. Perhaps, however, the tide is beginning to change. There’s a new game in town, sneaking up through the ranks of open source software, and it has the potential to turn VDI on its head.
The history of OpenStack
The best ideas of all, often spring up from Pad Thai. If that’s not an ancient proverb, perhaps it should be. In the summer of 2010, engineers from NASA and Rackspace met at a Thai restaurant near NASA Ames Research Center to discuss complimentary projects that were underway at their respective organizations. NASA was knee deep in the Nebula project, which included an open source platform called Nova. Rackspace was building a storage platform called Swift. By the end of dinner, the two parties resolved to combine their two projects into a single open source collaboration named OpenStack.
(Want to read the whole history? Wired published a fascinating read back in April 2012.)
October 2010 saw the first release of the OpenStack Cloud Computing Platform and, within a year, OpenStack was backed by big names like Dell, HP, and Cisco. Over the years, OpenStack has expanded from the original Nova and Swift projects to include networking (Neutron), block storage (Cinder), a dashboard (Horizon) and much, much more. But, where are the bits that enable VDI?
In 2012, ZDnet hinted that an open source consortia would soon announce development plans for an OpenStack-based VDI solution. That consortium, however, never seemed to materialize. Perhaps, in 2012, OpenStack was too new to branch out of its comfort zone as an IaaS open source cloud computing platform.
Now, however, the bits are all mostly in place. The only piece that’s missing in OpenStack, as pointed out in their documentation, is the connection broker. To manage VDI, or desktops-as-a-service, in OpenStack, you need to find a third-party connection broker to manage user assignments and connections to your desktops, and potentially a third-party display protocol to use for the connection. (In a previous blog, I covered features you should look for in a third-party tools for OpenStack VDI.)
The future of OpenStack VDI
If OpenStack VDI means gathering bits from a number of sources, why would anyone even consider it?
First, consider the flexibility that gathering those bits provides. You can choose from a number of hypervisor, including free or open source versions. You can select the appropriate connection broker and display protocol. You can manage both Windows and Linux operating systems. Sure, full-stack virtualization solutions simplify your choices, but that’s not always a good thing. OpenStack allows you to pick the best-in-breed component for each use case, and you can mix and match those components as required.
Second, consider the licensing fees VMware and Citrix require to implement VDI in their stacks. The various OpenStack projects provide, for free, functionality that VMware and Citrix license, not just for brokering connections to desktops, but for desktop provisioning, as well. The prospect of implementing VDI without expensive licensing is destined to open up VDI discussions in organizations that may otherwise have decided to do without.
Where once there were two
OpenStack continues to mature and the ecosystem around it continues to grow. To deploy a successful VDI using OpenStack, you need only find a connection broker and a display protocol that suits your needs. If an easy-to-implement, lower-cost blue print for such a solution becomes popular, OpenStack VDI could gain momentum. It may never take over the market from VMware and Citrix, but perhaps the duo could become three.