markrubens - Fotolia

Intel offers enterprises helping hand in making private clouds work for them

Chip manufacturer Intel sets out plans to help the enterprise deploy tens of thousands of cloud environments

Intel claims enterprises with at least 1,200 on-premise servers should have sufficient scale to make building a private cloud environment financially viable.

At the chipmaker’s Intel Cloud Day event in San Francisco, a slew of enterprises in the healthcare, financial services and education markets set out plans to build their own private clouds for business agility, data protection and cost-saving purposes.

During the event’s keynote, Diane Bryant, senior vice-president and general manager of Intel’s datacentre group, said enterprise interest in the private cloud is gathering pace, as firms look to make use of a mix of on- and off-premise environments dictated by the needs of the business.

“The dependency businesses have on technology continues to grow. That then drives traditional enterprises to deploy private clouds, both to get greater efficiency and agility for their existing services, but to also drive new revenue,” she said.

The long-term future of the private cloud concept has been the subject of much industry debate over the years. Market watchers ponder whether or not enterprises will want to build their own on-premise environments when using public cloud could work out easier and cheaper.

On this point, Bryant said private cloud can make good economic sense for enterprises already operating a sizeable server farm footprint.

“Our estimate is if you have between 1,200 to 1,500 servers under management, you have sufficient scale to deliver a very efficient private cloud system,” she said.

Private cloud use cases

US-based book publisher Houghton Mifflin Harcourt provides educational content for both school children and adults worldwide. It is using private cloud to deliver on its goal of providing people with “adaptive learning” experiences.

For example, by dynamically tailoring the way content is delivered to users based on their preferred learning styles and ability in near real-time – and that requires capabilities best delivered using the private cloud.

“The adaptive model requires an analytic processing in the background to drive that dynamic content,” said John O’Connell, the company’s senior vice-president of alliances and partnerships, at the event. “Ensuring student data privacy is central to our business.”

Cloud for all

A large portion of the event focused on the work Intel and its supplier partners are doing to make it easier for users to deploy private clouds, and help enterprises adopt a multi-cloud strategy involving service providers that span on- and off-premise environments.

One of the barriers to achieving this goal is the lack of interoperability between the cloud service providers and their environments. This is something Intel has been working to address for some time through its backing of the OpenStack platform.

To date, this has seen it team up with managed cloud service provider Rackspace to create an OpenStack Innovation Centre in San Antonio, Texas. The site is equipped with a 1,000 node compute cluster that developers can use to test and validate their cloud management offerings.

The pair, along with Mirantis, have also created the largest community of OpenStack developers in the world, claimed Bryant.

Together, they are contributing code to help increase the enterprise appeal of the OpenStack platform through the roll-out of features that permit live migrations of virtual machines and improve the performance of data analytics tasks, for example.

Read more about Intel

These initiatives are geared towards helping “tens of thousands” of cloud environments come online, and help cut the time it takes to deploy one to “less than a day”, said Bryant.

“We have been working with Mirantis, Microsoft, CoreOS, HP and VMware. We’ve made a lot of progress against this goal,” she continued.

“Collectively, we’ve published 20 reference architectures and demonstrated the success of those through proof of concept deployments with the users. We’ve got several of those in production deployments with users as well.

“We’re happy to say that several of those system stacks can go from rolling the rack of serves into the datacentre to having a user requesting cloud capacity in less than one day,” she added.

Read more on Datacentre performance troubleshooting, monitoring and optimisation