Inside Advanced Scale Challenges|Friday, November 17, 2017
  • Subscribe to EnterpriseTech Weekly Updates: Subscribe by email

OpenStack Backers Come Out Swinging 

OpenStack Summit

The OpenStack community is highlighting the cloud platform's steady progress within large enterprises during its annual summit this week in Austin, Texas.

Along with a growing list of hyper-scale use cases, the nascent movement sought to dispel deployment and complexity worries by claiming that half of Fortune 100 companies are now running OpenStack. A user survey released earlier this month found that 65 percent of OpenStack deployments are in production, up from 33 percent last year, and that 97 percent of users surveyed want a standard cloud platform and APIs.

Those statistics represent an effort to forge the establishment of a standard platform for building and scaling private clouds while providing cloud-native frameworks that would run on top of infrastructure and application "primitives."

"Being able to tie that all together is really, really valuable," stressed Jonathan Bryce, executive director of the OpenStack Foundation.

At the same time, Bryce stressed OpenStack's role in handling greater IT "diversity" in which existing enterprise applications must be maintained as new ones are developed and rolled out on increasingly shorter timeframes. That diversity also includes a growing number of use cases ranging from big data applications to network functions virtualization (NFV).

Indeed, Bryce noted, telecommunications and NFV are the fastest growing OpenStack use cases as carriers hustle to keep pace with the explosive growth in network traffic that according to an AT&T (NYSE: T) engineer at the summit is running an estimated 114 petabytes a day.

Sorabh Saxena, who oversees AT&T's software development and engineering operations, noted that network traffic is expected to jump ten-fold through 2020. Hence, the telecom giant was an early adopter of OpenStack as it deploys an integrated cloud service. OpenStack underpins AT&T's plan to shift 75 percent of its network infrastructure to the cloud while making greater use of software-defined networking (SDN).

Saxena said AT&T's cloud includes ten OpenStack projects with three more planned by the end of the year. The telecom giant's requirement is to reduce deployment times for cloud "zones" from months to days.

The carrier used OpenStack tools to develop a "resource manager" that among other things links network operators and tools to the AT&T cloud zones. In one example, Saxena said an SDN-based video service was rolled out to provide customers with an interface that allows them to reconfigure or make other changes to their networks without waiting for their service provider. The self-service capability includes a "local control plane" and APIs to manage networks.

Saxena described the OpenStack-based "network on demand" as the equivalent of "giving the customers the keys."

OpenStack proponents also stressed the growing need to automate the process of rolling out private clouds. Saxena drew applause when he announced that AT&T managed to deploy 54 cloud zones in two months.

These and other uses cases described at this year's packed summit are designed in part to overcome the perception that OpenStack is not ready for primetime. Boris Renski, co-founder of pure play OpenStack vendor Mirantis, also drew cheers when he noted in a keynote that early skeptics like market researcher Gartner have come around to the view that "OpenStack seems to not be crap!"

Renski argued that the key to OpenStack's success is "one part technology and nine parts people and process." That latter, he noted, is "boring," yet critical to successful deployments. The vendor has positioned itself "in between" customers who would "give up everything" in a shift to public clouds or doing everything themselves on a private cloud platform.

Mirantis and other OpenStack developers are addressing dull but critical enterprise culture and process issues while de-emphasizing technology. Ultimately, emerging cloud-native frameworks like Kubernetes and Apache Mesos cluster managers run on top of infrastructure and applications. Users won't be concerned with the underlying IT infrastructure, insisted OpenStack director Bryce.

About the author: George Leopold

George Leopold has written about science and technology for more than 25 years, focusing on electronics and aerospace technology. He previously served as Executive Editor for Electronic Engineering Times.

Add a Comment

Share This