Canonical fluffs one-click Ubuntu cloud stack
Linux, NoSQL, Hadoop – check
Canonical is accelerating Ubuntu's push into the cloud, delivering an integrated stack of cloud platforms ready for download.
Canonical has revealed that it's working with open-source project Hadoop and NoSQL database providers to deepen the level of integration between these big-data technologies and the Linux distro's next planned release: Maverick Meerkat, aka Ubuntu 10.10, due in October.
Separately, sources close to Canonical have told The Reg that the company is in talks with Cassandra and CouchDB on NoSQL, and start-up PuppetLabs for data-center automation and provisioning.
Canonical also has an internal project underway codenamed Ensemble to manage dependencies, deployment, and provisioning of applications on Ubuntu clouds.
Data center and application tooling will come in the follow-on Ubuntu 11.04 and 11.10 releases.
Canonical's end goal seems to be single-click deployment for a cloud from the Ubuntu command line, starting with specific applications.
Neil Levine, vice president of Canonical's corporate services division, said in an interview here with Dell's cloud-computing evangelist Barton George that Ubuntu is moving up the stack and bringing the earlier work on integration between Ubuntu and Eucalyptus to other applications for "deeper integration and slickness" in Ubuntu 10.10.
Canonical is targeting Hadoop and NoSQL – used by hyperscale providers like Yahoo! and Facebook – believing ordinary businesses are now ready to start use them for data processing and analytics.
"Business is beginning to use this kind of stuff now," Levine said. "We are trying to integrate those in [Cassandra and Hadoop] so you can run those kinds of applications on top of the public cloud using Amazon or in private cloud using UEC [Ubuntu Enterprise Cloud], so you can get easy access to those applications and plug them into the problem you have."
The push tightens Ubuntu's embrace of cloud. Building on initial support for Amazon Machine Images (AMIs) that contain applications, libraries, data, and configuration, Ubuntu has bundled Eucalyptus for Amazon-style elastic compute and storage on the server.
The focus on frameworks, applications, and application management means Ubuntu is buying into the concept of deploy-anywhere cloud computing.
Unlike VMware with vSphere and chief Linux rival Red Hat using KVM, Ubuntu's cloud stack does not seem to be anchoring itself in a virtualization layer that can tie in apps.
Instead, Ubuntu's focus is on ease of deployment, whether that's on a public Amazon cloud, private Amazon-like cloud, or on a mixed server infrastructure that does include virtualization from VMware and Red Hat.
Despite the hype over its ability to scale and its number-crunching potential, Eucalyptus is difficult for ordinary users to set up and rollout. By integrating Eucalyptus with Ubuntu, Canonical hopes to solve this for Ubuntu clouds.
A further plus of integration, meanwhile, would be that Canonical can provide customers commercial support for Hadoop, Cassandra, and CouchDB. ®