OpenStack platform has first major release, named Austin

The open-standards and open-source cloud project OpenStack has released Austin, the first version of its production platform.

OpenStack Austin is a cloud platform built on code from Rackspace's cloud files and cloud server systems, and from Nasa's Nebula Cloud Computing platform. Work on the cloud platform was initially announced in July, and the release of Austin came after developers added in support for more hypervisors and alternative cloud services.

Austin is the "culmination of what we set out to do months ago", OpenStack's head of business development Mark Collier told ZDNet UK on Thursday. He said the key new features for the release were the support of multiple hypervisors and the publication of an official OpenStack application processing interface (API).

OpenStack now supports the Linux-based Kernel Virtual Machine (KVM) and the Citrix-owned XenServer hypervisors, along with User-mode Linux (UML), which modifies the Linux kernel to give a hypervisor-like functionality. When asked why there is no support for other hypervisors, such as VMware's models, Collier said: "We would like to support VMware... we have a ways to go to enable that and that full compatibility." Support may come along in six to nine months, he added. Regarding Microsoft's hypervisor, Hyper-V, the release document for Austin notes that support for it "has been started and is scheduled to be included" in one of the next major releases.

The release also increases the portability and interoperability of cloud installations on the open-source private cloud Eucalyptus platform and on Amazon's Elastic Compute Cloud (EC2), by adding compatibility within OpenStack for handling security groups. This is a profile-based network firewall system for instances that are used by both of those clouds. This development makes it easier to migrate cloud installations on either of the platforms into OpenStack, while preserving the security element, Collier told ZDNet UK.

Lowering the pain of migrating away from Amazon Web Services is seen as a key way for OpenStack to boost its user base, according to software industry analyst Michael Cote of RedMonk.

Cote told ZDNet UK on Thursday that "for OpenStack — or any would-be public cloud — making migrations from Amazon easy is key as that's where the bulk of current cloud users are that you'd want to poach". "More generally, if there's a lower-level, architectural concept like security groups that's finding success in the cloud world, it's probably wise to implement it instead of coming up with your own ideas that developers will have to learn afresh," he added.

Further new features in the release allow greater granularity of statistics for instances within an OpenStack cloud, a rescue recovery mode to fix system and configuration errors in virtual machines, user-definable metadata for individual OpenStack user accounts, a way to set access controls and a partial implementation of a function for a virtual machine image registry service.

"Most people working in the cloud space, nascent as it is, are very interested in OpenStack," Cote said. "[Although] aside from Rackspace themselves, there are not big instances of using OpenStack that I know of. That's because OpenStack as a 'cloud in a box' platform is still a work in progress, requiring the deployer to customise it to their set-up."

However, a principle analyst for Analysys Mason, Steve Hilton, told ZDNet UK on Thursday that for many companies an open-source cloud may be a step too far. "There are a lot of different types of enterprises in the world and one size doesn't fit all, but there are plenty of large-size and small-size enterprise customers who would never do an open-source solution because they fear security issues... or they lack the internal coding resources to build on the platform and maintain it," he said.

Thirty-seven companies are now involved in the OpenStack scheme, including Intel, AMD, Citrix, Rackspace, Nasa, Dell, NTT Data, Rightscale, Scalr, Cloudscaling and InterNap.