It’s easy to forget network functions virtualization (NFV) is only one year old.
The concept was introduced at last year’s SDN & OpenFlow World Congress in Darmstadt, Germany. One year later, at the same show (this time in Bad Homburg, Germany), NFV is a headlining topic and an imperative for vendors and carriers alike.
It’s been a busy 12 months, as reflected by the status update that the ETSI NFV industry specification group (ISG) released this week. The group now counts 152 members, and NFV has become a major subject of discussion alongside software-defined networking (SDN).
NFV is about moving more network functions into software that runs on industry-standard hardware. (It’s not the same as network virtualization, which is about creating network connections on the fly. For some more detail, see NFV vs. SDN: What’s the Difference?)
The driving force for all this? Envy.
“Somewhere out there, there’s a creative ecosystem and I think it’s no more apparent than when you look at how many apps have been developed on top of Android and on top of Apple iOS,” said Don Clark, head of network evolution for BT Group and chair of the ISG’s network operator council. (That’s him in the photo above.) “I envy that dynamic of innovation. What NFV brings us is the ability to innovate the network in timescales unimagined.”
Getting NFV Down on Paper
This week, the NFV ISG released five documents:
- An NFV use cases document.
- A set of NFV requirements.
- An NFV architectural framework, which includes management and orchestration. This framework will handle virtualization-specific management tasks and is the interface to carriers’ OSS/BSS systems.
- A common terminology for NFV.
- Proof-of-concept guidelines — rules, in a sense, for submitting proofs-of-concept to be demonstrated at ISG members’ labs. This could go a long way in encouraging vendors and community members to devise creative uses of NFV.
The documents are not set in stone; they’ll be augmented, the white paper promises, with new versions likely to come in late 2014.
Clark described the documents in a Thursday morning talk at the SDN Congress. He also took the opportunity to emphasize the NFV group’s commitment to open-source principles.
The NFV ISG also releases a white paper announcing the five documents. (A sixth document!) It notes that carriers will have to rethink their OSSs, which Axel Clauberg of Deutsche Telekom described in his keynote as “the main contributor to the long lead times we have.” Carriers will have to decide whether to change OSSs in a step-function manner (ripping off the Band-Aid, so to speak) or gradually.
“For those on the step-change path, NFV could provide an opportunity to leverage NFV management and orchestration to transform the current OSS into a more efficient system. For those who wish to take the incremental path, NFV can also be introduced in a way that minimizes the impact on existing OSS and operations models,” the white paper reads.
Translating NFV Into Euros
The SDN Congress’ NFV track ended Thursday afternoon with a panel of carriers discussing where the technology goes next. It was a mix of excitement and trepidation, as carriers are beginning to grapple with the reality of actually implementing these ideas.
Not that that’s a struggle in every case. The enthusiasm at DT is overflowing, said Franz Seiser, VP Core Network and Services, Europe & Technology at Deutsche Telekom. “Maybe I am a good sales guy; I don’t know. But all of them understand the principle. All of them want to do something,” he said. “I really have a challenge: ‘Guys, don’t be too fast. The technology’s not out there!”
Others are more conservative, as you might expect. “We need to make a business case that can be understood” by the non-technical parts of the company, said Yves Bellégo, director of network technical strategy for Orange Group. “We say it will improve some [customer] experience, gain some flexibility — but in the end, we have to translate that into euros or dollars.”
Don’t assume, though, that NFV will languish the maze of carrier bureaucracy. The envy that Clark mentioned is permeating the telecom industry, and panelists revealed an eagerness to shed that image of big, slow carriers. Cloud providers and over-the-top services have shown them a world where not everything has to be built to the ironclad standards of the old telephone network. Carriers have to be willing to accept that things will fail, and that services can survive those failures, said Tetsuya Nakamura, senior research engineer for NTT Docomo.
Life With Less Than Five Nines
Failure recovery is something that “needs to be changed by software mechanisms and not by having five nines on every box,” Michel said, referring to the infamous 99.999-percent reliability required of old telco hardware.
None of this means carriers want to get careless, though. It’s imperative that all this newfound flexibility and willingness to fail doesn’t affect the quality of the services. “For the customer, nothing must change,” Michel said.
Behind all this is the fact that NFV is in its infancy. One theme throughout this week’s Congress was that carriers are eager to play with this new toy and find out how it might change their lives.
“We need to go through the steps. We need to go through real field operations so we can get more experience from NFV,” Bellégo said. “We do not expect to get that in even three or five years, but we are on the right track.”
Did he literally mean three or even five years? Possibly not; some carriers seem anxious to move faster than that — but whatever the timeline, they seem to agree about being on the right track. NFV might not solve their problems, but it seems poised to make a go of it.
See other coverage of the SDN & OpenFlow World Congress: