The transition to NFV networks isn’t happening fast enough for one of the thought leaders involved in the original 2012 white paper that operators released, but moves are afoot that could change the game over the next few months.
Don Clarke, a 25-plus-year R&D veteran of British operator BT, was the editor of the joint carrier white paper that ignited the NFV space about five years ago. (Margaret Chiosi, who was at AT&T at the time, and Prodip Sen, then at Verizon, were also part of that initial group.) In 2014, Clarke joined CableLabs, where he is principal architect of network technologies, and continues to work closely in the NFV/SDN space.
“It’s taken longer than I’d hoped” for the NFV space to mature, he told FierceWirelessTech. But having said that, he admits that throughout his R&D career, he’s tended to underestimate how long it takes for any new innovation. And he pointed out that the authors wanted to be careful to avoid hype.
As part of its NFV Release 2 program, ETSI is preparing to publish key specifications that are expected to have a profound impact in terms of making NFV interoperable. These specifications include VNF Packaging and Protocol and Data model specifications—APIs, around the ETSI NFV Management & Orchestration Framework (MANO). The open source community has been kind of working independently, but standards are important in the telco/cable world due to network scale and longevity as well as regulatory requirements.
When it comes to connecting the last mile, cable companies have a very different network than the telcos. When Clarke joined CableLabs, he was very much aware that he was joining an industry with a very powerful high-speed ubiquitous access network, and that’s where the key differences lie. Outside of that, everything else is going to look similar.
“That’s why I’ve been very active in international standards working with telcos because the components that we’re going to use to assemble these networks in the future are absolutely common,” he said. “We want to use the same data center compute technology as the telcos do and we want to use the same core network and basic fundamental network capabilities and standards as they do, but where we’re going to see differentiation is definitely where we use the access network and in particular when we overlay things like wireless on top of that. Then we’re going to use NFV and SDN in specific ways. The stuff that we’ll use to upgrade those networks will be the same kinds of things that the telcos use; we’ll just put them together in a different way.”
In fact, when it comes to NFV, the cable industry is very keen to make it happen and just as committed as the wireless industry, according to Tetsuya Nakamura, principal systems architect for core innovation at CableLabs.
But there are a lot of reasons things are not moving as fast as originally thought. Back in the early part of 2013 when the NFV/SDN efforts were kicking off for telco, “we thought that all we needed to do was to converge requirements and the industry would go on and do it,” Clarke said. But that hasn't been the case, in part due to complexity, partly because people didn’t see clearly how this stuff would work and partly because the skill sets in the industry are not well aligned. Telco engineers tend to be very good at networks and not good at software, and that’s still a bit of a quandary.
“You really do need hybrid skill sets, and that’s a really difficult thing for the operators to get their head around,” Clarke said. They’ve got a legacy skill set that has been fine-tuned to deliver and operate networks at vast scale, but these networks are no longer going to be designed the way they used to be.
“They’re going to be designed with software, and that’s the world of the Googles and the Amazons, so we’ve got to try and get skills to align with this new game,” he said. That’s part of the reason it’s taken a bit longer.
But as ETSI releases new specifications this year, those will be available for the open source communities to adopt, and fragmentation in the open source world may sort itself out. That’s the hope anyway.
Clarke added that it’s not just ETSI—other entities, like MEF and the Broadband Forum, are contributing as well, and all these efforts will be necessary to pull off the kind of transformation the telco, cable and wireless industries are talking about.