As usual, this release includes updates to all of the platform’s services, which now span the gamut from cloud computing to storage services, identity management and tools for processing large amounts of data. The highlights of today’s release, however, are the addition of a new data processing service that automates the provisioning of big data clusters, the addition of more granular policy controls for Object Storage and initial support for network function virtualization (NFV), something OpenStack’s users in the telco industry have been asking for.
As the OpenStack Foundation’s COO Mark Collier told me earlier this week, this release clearly shows the fingerprints of OpenStack’s users and operators. The prime example for this is probably NFV. “I believe this is a very large opportunity for OpenStack,”Collier told me. With NFV, telcos — but also large enterprises that run their own networks — can move many of their networking services from expensive proprietary hardware to commodity servers. That comes with huge potential cost savings, but the software stack also has to be extremely stable and offer real-time performance.
“If I’m a telco and want to move a hardware system for voice calls to software,”Collier noted, “I need to know that this vitalization layer won’t start dropping calls.” In this initial release, then, the team focused mostly on performance and will continue to offer more functionality in upcoming releases.
The main new feature of the OpenStack Juno release is the launch of the platform’s new data processing service. This automates the provisioning and management of Hadoop and Spark clusters for big data analytics. The service started out as a straight Apache Hadoop service, but over time, the developers also added support for HortonWorks’ and Cloudera’s Hadoop, as well as Spark.
In our interview, Collier also noted that he believes the addition of more granular policies for OpenStack’s Object Storage as an important update for many users. With this, users get more control over how they want to store, replicate and access data across backends and regions. Often, for example, you have data that you can easily replicate, so you don’t need to have three backups available at all times.
With this update, OpenStack is also launching federated authentication across OpenStack clouds. OpenStack users like CERN, for example, have their own private clouds and then augment them with capacity on public services. Now, they can give a researcher a single set of credentials that gives them access to both.
In total, the Juno release features 310 new features and 3,200 bug fixes. You can find a more detailed set of release notes here.