The latest milestone release of OpenStack, the open source cloud platform project that's become a centerpiece of both Red Hat's and Canonical's cloud-building efforts, is out. But as advanced as this release is over its predecessors, a few key integrations for OpenStack remain as future to-do items -- and there's tension between the pace of OpenStack releases and the curve of its overall adoption.
Code-named "Icehouse," the newest OpenStack release (ninth in all) comes fairly shortly after the previous major release, "Havana," which dropped in October 2013. That version added major changes to OpenStack's networking, orchestration, and metering systems for better functionality to the service providers who have become some of OpenStack's most vaunted adopters.
The base of Icehouse contributors has grown 32 percent since Havana, and many of the changes "[reflect] a community-wide effort to bring the voice of the user into [OpenStack]," according to the company. Among the new changes this time around that are vaunted as being user-centric:
Live upgrades: With previous editions of OpenStack, an entire OpenStack cloud would have to be shut down in order to roll out upgrades. Icehouse now makes it possible to "upgrade controller infrastructure first, and subsequently upgrade individual compute nodes without requiring downtime of the entire cloud to complete," according to the OpenStack wiki. The current iteration of this feature provides live upgrades in a limited way, but the migration process was a commonly cited stumbling block for OpenStack adoption; having some amelioration for it is a positive step.
Federated identity: Multiple OpenStack nodes can now be logged into with a single ID provided through the federated identity system Shibboleth. This was a feature requested by none other than Europe's physics research institute CERN, which switched from OpenNebula to OpenStack as of 2013.
Trove: Originally known as "Project Red Dwarf," Trove is a new OpenStack component designed to allow management of database resources. MySQL, for instance, can be managed through Trove so that MySQL users and schemas can be manipulated via Trove's API. Support for other database technologies is on the way as well, including NoSQL systems like MongoDB, Couchbase, and Cassandra.
Object storage replication: OpenStack's Swift object storage system has a new replication mechanism, called ssync, that intercepts requests coming in and out of Swift and syncing in a more intelligent manner than the existing rsync-powered mechanism.
Most of OpenStack's greatest challenges are a matter of the market rather than technology; specifically, it's having difficulty finding adoption on the same level as Amazon or other cloud services. Various success stories continue to pour in -- Red Hat recently announced a slew of major OpenStack wins with various companies and educational institutions -- but the product still shows its strongest uptake with those who are already committed to supporting and developing it, and not in enterprises generally. The question isn't whether OpenStack can grow into that over time, but rather whether it can do so before public cloud providers become the standard.
Another ongoing challenge for OpenStack -- one that won't be solved directly by the introduction of new features -- is the challenge posed by economies of scale when competing against public clouds. NASA dropped out of OpenStack after conducting its own assay of the cost effectiveness of the project and found public clouds were a better deal. The most recent round of price slashing and feature packing by three of the biggest cloud vendors shows they're determined to remain in both the public and private cloud game, at least in terms of pure value for money.
None of this seems to have deterred the rush to pack more features into OpenStack, though. The "Juno" release, set for later this year, will roll in possibly the biggest and most influential addition to OpenStack yet: support for Hadoop by way of the Sahara project. The folks at OpenStack will have their work cut out for them, given how many other companies are now offering Hadoop as a feature-complete data management solution and not simply as an element to be orchestrated.
This story, "OpenStack's latest release keeps the DIY private cloud features coming," was originally published at InfoWorld.com. Get the first word on what the important tech news really means with the InfoWorld Tech Watch blog. For the latest developments in business technology news, follow InfoWorld.com on Twitter.