The open source OpenStack cloud platform has been widely deployed around the world by network operators and service providers alike.
OpenStack isn’t just about enabling cloud compute utility computing; it has also become a preferred platform for enabling and deploying Virtual Network Functions (VNFs). In a session at the OpenStack Summit Berlin, Henrik Blixt, product line manager at VMware, outlined some of the key attributes that have attracted service providers to OpenStack. Blixt also provided guidance on some of the key challenges of OpenStack as well.
So why are operators choosing OpenStack?
Breaking Free from Legacy Hardware
For most the history of the networking business, hardware and software were tightly coupled, and that was especially true for service provider networking. Blixt commented that specialized hardware also meant tight lock-in with vendors, as well as somewhat inflexible solutions.
He also noted that upgrading hardware is often complex, expensive and time consuming. Additionally, the legacy skills required to maintain proprietary hardware has become increasingly expensive.
“That’s one of the driving forces of the move to VNFs, is getting off legacy hardware and moving to something that is easier to manage,” Blixt said.
Moving to a cloud-native approach, isn’t just about software-defined networking (SDN) or repackaging code in containers. The promise of cloud-native technology and OpenStack as a platform is that it can enable operators to innovate faster, according to Blixt.
While cloud-native technologies and process offer a more agile and flexible approach than legacy hardware systems, cloud native also introduces new challenges. Blixt said that for some organization implementing new processes can be more difficult than deploying new hardware.
He added that cloud native also introduces another toolset and set of technologies that organizations need to test and train staff on how to operate properly. Additionally, the fast pace of cloud-native technology is another potential challenge.
“Investing and committing can be risky and costly in areas with no clear market leader,” Blixt said.
Building a Software-Based Carrier-Grade Environment
The old model of legacy hardware included certain capabilities and guarantees that are now being replicated in software models like OpenStack.
The primary step to building a software-based carrier-grade environment is having some form of cohesive configuration management approach. Blixt warned that having too many disparate tools with too many exposed knobs is a potential risk.
Having all the right software dependencies in place is another key attribute, as is having resource guarantees. Overall, the “carrier-grade” label implies a high level of resiliency and fault tolerance that now needs to be replicated in software.
“A lot of network services are mission-critical. You can’t dial 911 and not be able to access the network,” Blixt said.
The Promise of OpenStack
Among the big attractions for operators when looking at OpenStack is the community of participation. Blixt noted that most of the world’s major service providers now use OpenStack somewhere in their infrastructure.
Having so many operators working and using an open source technology enables cross-vendor collaboration.
“Vendors focus on solving problems, not building proprietary solutions,” Blixt said. “Sharing ideas and working together fuels innovation.”
Sean Michael Kerner is a senior editor at EnterpriseNetworkingPlanet and InternetNews.com. Follow him on Twitter @TechJournalist.