Server NewsOpenStack Diablo Open Source Cloud Platform Accelerates High-Availability

OpenStack Diablo Open Source Cloud Platform Accelerates High-Availability

ServerWatch content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

After six months of development, the OpenStack Diablo release is now available. OpenStack is an open source cloud platform originally started by Rackspace and NASA that has grown to more than 80 member companies.

The Diablo release is the first major release of OpenStack since the Cactus release in October. The new Diablo release includes a long list of new features, including a high-availability mode, a distributed scheduler and data synchronization capabilities. Diablo is also showing off some technical previews for new networking capabilities, identity management and a cloud management dashboard.

“Diablo was our first six month release cycle, the other releases were all on a three month cycle,” Jonathan Bryce, chairman of the Project Policy Board for OpenStack told

Bryce explained that the longer release cycle has led to many feature improvements. Among the improvements is a high-availability controller, which is a change from the previous failover configuration.

“So you can distribute the network manager across all of the compute nodes, which gives you higher availability,” Bryce said.

On the scalability side, Bryce noted that there is a new distributed scheduler that allows administrators to provision machines across multiple zones and clusters. Bryce explained that there were a few schedulers that were already present in OpenStack, but they were fairly basic in his view.

“They were not aware of multiple clusters,” Bryce said. “When you want to be able to control virtual machines across multiple zones, you have to handle provisioning differently.”

Bryce added that the scheduling algorithms will continue to evolve in future OpenStack release to further add capabilities.

From an object storage perspective, Diablo enables data synchronization across multiple clusters.

“This is a great feature if you want to have guaranteed disaster recovery where you have data that you absolutely cannot lose,” Bryce said. “You can set this up so you can replicate on a container-by-container basis between object storage clusters.”

The other big thing that went into Diablo is a number of incubated projects that will become core projects in the next major release of OpenStack. One such incubated project is the Dashboard project, a web-based interface for managing resources in OpenStack clusters.

Bryce explained that the dashboard has both a user view and an administrator view. As an administrator, you can create users and provide different kinds of limits on resources; as a user you can start up virtual machines and control storage.

The other incubated project is the Keystone identity management service, which helps integrate an OpenStack deployment with an existing authentication system.

“So if you have ActiveDirectory you can use the OpenStack identity management service to tie in your credentials that your employees already have and enable them to provision resource across the cloud,” Bryce said.

Diablo also provides a preview of the Project Quantum Networking project for OpenStack. Quantum provides more advanced networking options that enable OpenStack to integrate with existing networking gear and network policies for quality of service.

One area that remains a work in progress is enabling migration across different OpenStack cloud implementations. A potential solution for that dilemma is already in Diablo with the Glance project, an image service that debuted in the Bexar release of OpenStack in February.

Bryce said there have been a number of Glance improvements as part of the Diablo release.

“One of the big improvements is that the API now supports searching and filtering when you are querying for images,” Bryce said. “There has also been a decent amount of work in figuring out how move the image files around in a cluster.”

While OpenStack cloud portability is still somewhat of a challenge, Bryce said you can run Glance and take an image snapshot and then spin it up on another OpenStack cloud.

“There is also some work that is being done to establish a reference architecture for the hardware and the image platform that will allow you to have a baseline for moving across different OpenStack clouds,” Bryce said.

The other key thing being worked on is something known as the faithful implementation test suite.

“That suite is set of tests that you can run against an OpenStack instance to verify that it has the proper APIs and it support the standards image and hypervisor formats,” Bryce said. “It will allow clouds to be certified as OpenStack and that interoperability between them is possible.”

Bryce noted that work on interoperability is part of the next development cycle, which is now kicking off. The next major release of OpenStack is codenamed Essex. It is set for release in six months.

Sean Michael Kerner is a senior editor at, the news service of, the network for technology professionals.

Follow ServerWatch on Twitter

Get the Free Newsletter!

Subscribe to Daily Tech Insider for top news, trends & analysis

Latest Posts

Related Stories