What’s New in OpenStack Grizzly

If you’ve been following the OpenStack community, you know that the last two weeks have been busy. The Grizzly release reached general availability on April 12, and was immediately followed by a flurry of presentations and hype at the OpenStack Summit.

Grizzly touts 230 new features, and the summit made it clear that the developer and adopter communities are growing rapidly. This begs the question: out of all the hype, what is truly significant about Grizzly?

Here’s my short list.

OpenStack now supports bare-metal provisioning.

Technically, this has always been possible, but Grizzly is the first release that treats bare metal as a legitimate target for cloud computing. I am pleased; I’ve long said that although virtualization is a key enabler of cloud, it is not its defining characteristic. By bringing physical hardware under the OpenStack umbrella, this release removes a stubborn barrier to pervasive adoption.

OpenStack is firmly on the SDN bandwagon.

OpenStack was already headed this direction, but Grizzly reinforces the trend by including support for more pieces of programmable network infrastructure, including Big Switch, PlugGrid, Brodcade, and Midonet. It now supports distributing certain network configuration duties such as DHCP across multiple servers, and makes load balancing an API-level operation. Many of these configuration choices are now exposed through Horizon (OpenStack’s dashboard).

OpenStack object storage is friendly to distributed consumers.

Grizzly includes support for cross-origin resource sharing (CORS), which allows browsers to interface directly with back-end storage without streaming through a mediating server on the back end. This allows far more efficient pairings of distributed object storage and the distributed clients that wish to use it.

OpenStack identity is maturing its multitenancy features.

Grizzly includes support for groups, impersonation, delegation, and other aspects of role-based access controls. Deployments that need to isolate departments or clients from one another will be pleased.

Of course, there are plenty of other goodies in the Grizzly release—but I think these will make the most difference, to the most people, over the longest horizon.

Adaptive Computing has been experimenting with OpenStack capabilities for a number of months. I am excited about what we have learned, and I look forward to telling you more about our findings in the near future.

Facebook Twitter Email