Why do telcos and OpenStack go so well together? What opportunities can telcos realize through an OpenStack implementation? And how does NFV, cloud and 5G fit into the telco business model? Red Hat's vice president of telecommunications, Darrell Jordan-Smith, joined Stu Miniman, senior analyst at Wikibon and John Troyer, chief reckoner at TechReckoning, in "The Cube" at OpenStack Summit this past May to talk telco and OpenStack.

In the interview, Jordan-Smith explains that telcos need to change the way they deliver services today. Telcos, he says, are searching for ways to modernize their network infrastructure to improve agility and to prepare for 5G, the next generation of services they'll deploy. They can't deploy 5G unless they have "an agile infrastructure fabric and an agile software capability to deliver those applications over those networks," he says.

Traditionally, telecommunications providers put hardware-specific appliances into network operations centers. "In the U.S. there are tens of thousands of them," says Jordan-Smith. However, today many telcos are moving to a software-based model where they don't necessarily need to buy a fixed appliance, with many choosing commercial off-the-shelf, x86-based technology deployed around an open compute platform. With this model, he says, "you're looking at storage, compute, networking in a scalable fashion, using open source technologies to deploy that in a massive scale."

Many of today's telcos are pushing services out to customers with software, and changing core network infrastructure to support that. Jordan-Smith pointed to the network edge, and increased agility in network operation centers at the edge, which enables telcos to push services out closer to the customer, as well as aggregate content and data that their customers require. "So, for example, say you take a video clip on your phone. There's no point in storing that in the core of the network, you'd want to store that at the edge where some of your friends would share it at that point in time." Today's telecommunication companies "want to make the edge more efficient, they want to build clouds around the edge, they want to aggregate all of those different clouds and they want to build agile-based infrastructure," says Jordan-Smith. He pointed to Facebook, Amazon, Netflix and Google as examples of service delivery that telcos are pursuing. "They need to get into that space in order to be agile enough to develop and deploy their next generation of applications and services."

"So, at this point, OpenStack in its evolution with this customer vertical, it seems like we're not only talking about a cloud, but maybe a cloud of clouds," suggested John Troyer, chief reckoner at TechReckoning, and co-host of the interview.

Absolutely, says Jordan-Smith. "[Telcos] typically have one of everything. They are looking at decoupled solutions in terms of the network-based infrastructure. They want to be able to manage every layer of that infrastructure independently of the other layers, in order to drive maximum flexibility and agility into their infrastructure, but also so they don't get locked in to any one vendor. That's a big, big theme in the telco space."

And because telcos are internally oriented toward engineering, and a "we'll build it and customers will come" perspective, many are going through a cultural shift with a renewed focus on building sticky services faster in order to attract and retain customers and reduce churn.

According to Jordan-Smith, telecommunication providers are driving innovation through open source. They want to create a flexible network infrastructure so the next Uber, Airbnb or WhatsApp can "develop innovative, next-generation content and services over their infrastructure" in an effort to make their services sticky for customers.

And what about implementation? Considering these big shifts in culture and technology, roll-out is a "crawl, walk, run process," according to Jordan-Smith, with operators looking at specific services and building slowly, service-upon-service.

Deployment takes time not only because open source very different from what telco organizations have used in the past, but also because the deployment process is now drastically changed due to market needs. "Traditionally they'll buy an application through an appliance, they'll take 9 months to deploy it in all their centers, and then another 3-6 months later they might switch it off. In the software-agile world they've got to condense that 12-to-18-month period down to maybe 3 or 4 weeks," he says.

Jordan-Smith also points to regulation as a driver of complexity. Because they are regulated companies, telcos must maintain quality of service throughout next-generation network deployment in order to bill and meet regulations in the markets they serve. This creates a very different scenario for telcos than it does for the Googles and Facebooks of the world, he says. "[Telcos] have a level of discipline that they need to achieve in terms of the availability of their network infrastructure, the availability of their services, the availability of their applications, and that links into a whole quality of service experience for their customers, and it links into their operations systems support, into their billing system, and the list goes on and on and on."

That said, since a lot of telcos are "engineeringly-orientated," as Jordan-Smith says, "it's great working with them because they really understand the difficulties. They also know that they want to build it and own it and understand it themselves because of their business model. To them, the network is an asset. It's not something that they can just outsource to someone else that doesn't necessarily understand that same degree of their asset."

"What we're seeing in Open Stack is not if but when [telcos] will deploy in mass scale," he continues. "And we're beginning no to see a general acceptance that this is a methodology and/or a technology that they can deploy and will deploy in the NFVi context. The other thing that's occurring in the space is they're looking at traditional IT workloads, so a telco-based cloud, if you want to use that terminology, is just as capable of running IT-based workloads and services as well."

Some telcos are looking to attract 5G and IoT where public cloud applications and services are being developed. According to Jordan-Smith, "they're looking at the insurance industry, the automotive industry, the manufacturing industry, and creating an environment where those applications can be built with many, many thousands of millions of devices connected to them. So, I think the definition in North America of public cloud infrastructure is going to evolve in that direction." He acknowledges, however, that the market differs in Latin America and Europe, where some telecommunications companies believe that they can be competitive in the broader cloud services space.

And how does 5G fit into the OpenStack picture? "5G is the reason why telcos are building NFVi," says Jordan-Smith, "because they realize that to connect all of those devices to their network-based infrastructure, they need to do it intelligently, they need to do it at the edge, and they need to have a high degree of flexibility and agility to their network-based infrastructure to create an innovation environment for application developers to connect all of those devices."

Jordan-Smith pointed to smart cars as an example around 5G, explaining that you need low latency, high availability, and reliability, "plus you need a portfolio of developers who are going to create all sorts of different applications for those vehicles that we're driving around on the street. Without 5G that does not happen."

Jordan-Smith wrapped up the conversation with some thoughts on the global reach of 5G. Asia is heavily invested in 5G based-infrastructure, and the operators there don't have a lot of legacy-based infrastructure that they have to replace. He points out that the applications and services and the way people experience the internet, or network-based access, in Asia is very different from the way users experience it in the U.S. or in Europe, which leads to different applications and business models evolving in various markets throughout the world.

With regard to the North American market, he sees a need to look at operators' differing business models to examine some of the higher order of applications and services that may be driving stickiness for the telco's own infrastructure and network services, as well as some of the more advanced applications like smart cars, smart homes, smart cities, energy, or better distribution methods for delivering products to the home.

You can watch Jordan-Smith's full interview here.