Cooking up some Standard Recipes for OpenStack with Opscode

The battle royal that is being played out between the different Cloud “operating systems” reached new heights at Structure last week. While all of us are talking about this stuff however, vendors and customers alike are simply trying to get going with these tools. They’re not focused on idealogical battles, but rather on finding easy ways of getting their cloud up and running. After all the tension of the last couple of weeks, it’s nice to see some deployment-focused announcements coming out.

Today it’s the turn of Opscode, creator of the Chef open-source systems integration framework built specifically for automating the cloud. Opscode is announcing the publication of Chef for OpenStack, a reference deployment for building and managing OpenStack powered clouds. The project has been sponsored by Intel and driven by contributions from a bunch of OpenStack players.

The product includes five cookbooks aimed to ease deployment and automation of computing, object storage, image services, dashboard and identity services of OpenStack Essex. Versions of the cookbooks will also be published for the Intel CloudBuilders community, a cross-industry initiative aimed at making it easier to build, enhance, and operate cloud infrastructure.

At the same time Opscode also today announced the completion of a new knife command-line plugin for OpenStack Essex, enabling users to create, bootstrap and manage OpenStack Essex compute instances. With Opscode Chef, OpenStack users can automate everything from server provisioning to application deployment in OpenStack-based cloud architectures directly from the command line.

The cookbooks being released today include;

  • OpenStack Compute: Enables enterprises and service providers to offer on-demand computing resources, by provisioning and managing large networks of virtual machines. Compute resources are accessible via APIs for developers building cloud applications and via web interfaces for administrators and users.
  • OpenStack Object Storage: Provides a fully distributed, API-accessible storage platform that can be integrated directly into applications or used for backup, archiving and data retention.
  • OpenStack Dashboard: Provides administrators and users a graphical interface to access, provision and automate cloud-based resources.
  • OpenStack Image Service: Provides discovery, registration and delivery services for disk and server images.
  • OpenStack Identify Service: Provides a central directory of users mapped to the OpenStack services they can access, acting as a common authentication system across the cloud operating system.

MyPOV

One of the major criticisms of OpenStack (see disclosure page) has been the fact that with so many different flavors of the product, it’s a very fragmented platform and reduced the reality of the “port your workloads across vendors” story. Any vendor that is actually able to install a consistent flavor of OpenStack across different vendors is on to a good thing. The jury remains out however on how possible that really is, in talking with enStratus’ George Reese about the topic I asked him whether, in his view, an automated deployment tool would really work across vendors – he was fairly doubtful suggesting that the differences between the various distributions are just too great to enable a single deployment tool across vendors anytime soon, if ever.

For their own product, enStratus “tunes” their tool for each specific vendor’s flavor of OpenStack. That task is much more challenging for OpsCode since the installs are so different across tools. I’m not suggesting that OpsCode can’t do at least some of what it says, but I’d suggest that the day that we’ll see painless cross-vendor OpenStack installation in an automated fashion is a fair way off.

Enhanced by Zemanta

1 Comment
  • Pingback: Cooking up some Standard Recipes for OpenStack with Opscode … | cloudguys.net

  • Ben,

    Good read. I’m actually quite sorry we were too busy to make it to the Structure conference this year. You really did hit the nail on the head about OpenStack. But there was this tidbit: “Any vendor that is actually able to install a consistent flavor of OpenStack across different vendors is on to a good thing.”

    That is precisely what we have done with our IPSaaS technology. While we have elected to use the same partner in Rackspace given their track record we have done testing with multiple SAN’s & Hypervisors and can successfully say that we can in fact accomplish just that. OpenStack running on multiple vendors. Though I do agree we haven’t automated that across multiple vendors. Today we can only automate Rackspace, LiquidWeb & a few smaller IaaS vendors though the process shouldn’t be much different on the larger IaaS providers.

    Keep ‘em coming.

Leave a Reply