Ramping Up Our OpenStack Investment, Involvement

Hustling to release a full suite of open cloud products built on OpenStack last year reduced our overall involvement in the OpenStack community. It also created some implementation specifics that were out of sync with common practices in other OpenStack implementations.

During my presentation at OpenStack Summit in San Diego last October – along with digging into some of the early success that we had – I committed to address these issues.

Our commitment to OpenStack is stronger than ever. As we head into OpenStack Summit Portland, I want to take a moment to update everyone on our progress.

One notable success for us included getting Cells merged into the Grizzly release of Nova. Cells is the level of host abstraction that will enable us to scale to tens of thousands of nodes behind a single API endpoint. Having it in Grizzly trunk eliminates the major difference between our deployment and trunk and also makes cells available to all OpenStack deployers. A big thanks to all who helped in making this happen.

We’re still plugging away on efforts to do continuous delivery from Nova trunk to our production environments. Our deploy pace has slowed as we are learning how to scale this effort and keep up with the pace of commits. It’s become clear that others in the community are interested in making this process work as well. So, we sponsored a work session here at Rackspace that included participation from Red Hat and HP. This is in preparation for a number of sessions at OpenStack Summit Portland on deployments and particularly CI/CD. To help facilitate progress, we’ve open-sourced our testing framework, Cloud Cafe, and engaged with the community to talk about how we might move our extensive test suite further upstream in the process. There is a lot to sort out in this area and we look forward to active collaboration across the teams to keep moving forward.

We are also taking active steps to align Rackspace with the rest of OpenStack around queuing, metrics and orchestration. At the Design Summit, we helped launch the Marconi project to focus on a queuing service optimized for OpenStack. Since the Summit, an initial community has formed around the project and the API is taking shape. The team looks forward to sharing its progress in Portland. Before knowing how OpenStack would evolve its focus, we had begun internal projects in both of these areas. We moved quickly after the Summit to align our metrics effort with Ceilometer. While we haven’t merged all of our work into the projects yet, we have our developers actively engaged and are far down that path. This week, we stated our plans to engage a couple of our teams in the Heat project to join forces with the community on orchestration and auto-scaling. These are exciting and vital project for the future of OpenStack and we’re pleased to join in these projects.

We also realize that the Rackspace public cloud has a number of implementation details that are out of step with common practice in the OpenStack community. Most of these were due to either our early lockdown on features to enable our 2012 launch – or decisions designed to keep in line with our first generation Cloud Servers offering. While  we believe some variation in implementations will be inevitable, we do want to eliminate as many of these as possible to provide as much of a common OpenStack experience as we can. Along those lines, we are committed to take the following steps between now and the end of 2013. Hopefully, this will close most of the gaps that may currently exist:

Nova API Extensions: Support for os_keypairs, config drive for file injection and eventual deprecation/removal of personalities, console/boot log availability

API Support: Exposure and full support of the Glance and Quantum V2 APIs as the primary interfaces for handling images and networks

Image Compatibility: Rackspace will begin supporting cloud-init (for Linux, and a Windows equivalent, likely cloudbase-init) style initialization, removing the dependency on the current Nova and Cloud Servers agents. In addition, we will focus our efforts on ConfigDrive and the Metadata service for client configuration to remove dependencies on XenStore and provide greater portability between implementations. The end goal is to enable images to be easily shared between the Rackspace Open Cloud and other OpenStack implementations with minimal effort.

We’re excited to be ramping up our investment and involvement in OpenStack. I hope this brief update gives you a sense of the energy we all feel around this great effort. If you have any questions or want to talk about these efforts further, find me in Portland!

Rack Blogger is our catchall blog byline, subbed in when a Racker author moves on, or used when we publish a guest post. You can email Rack Blogger at blog@rackspace.com.


  1. Hello,

    I appreciate your efforts to provide a common OpenStack experience for your cloud offering. We are a Rackspace customer (dedicated servers) and when we will move to the Cloud, OpenStack compatibility will be a key criteria for us.

    Best regards,


Please enter your comment!
Please enter your name here