Real-world challenges come with the territory, as NFV starts to see practical implementations.

Mitch Wagner, Executive Editor, Light Reading

October 5, 2015

10 Min Read
Getting to NFV Everywhere

One sure sign of maturing technology is that users talk less about blue-sky promises, and more about implementation challenges and how to solve them.

In other words, when people stop burbling excitedly and start sharing tips, you know that real transformation is afoot and benefits are coming.

That's true for NFV. Network operators are implementing NFV in real life, and vendors are shipping products. They've got the knuckle-scrapes and bruises to show for it.

The biggest experts on NFV came to Dallas from around the world recently to tell the NFV Anywhere conference what they learned using NFV in their transition to the New IP.

The conversation was a timely one, because the NFV market is booming. The global NFV market will grow from a base of $485.3 million in 2014 to be worth $2.2 billion globally by late 2015, and by 2019 generate business worth $12.7 billion, Heavy Reading analyst Jim Hodges said at the conference. Proofs-of-concept (PoCs) are moving into the commercialized phase this year, driving a significant increase in the adoption of NFV-related technology. An increase in spending on NFV projects is not linked to an increase in capex budgets, but is more about the reallocation of budgets away from traditional hardware and software.

Figure 1: State of the Market Heavy Reading analyst Jim Hodges discuses the state of the NFV market, while Hitachi's Paul Bourdreaux, Mitel's Kevin Summers, and TELUS's Nima Salehi listen in. Heavy Reading analyst Jim Hodges discuses the state of the NFV market, while Hitachi's Paul Bourdreaux, Mitel's Kevin Summers, and TELUS's Nima Salehi listen in.

So what were some of the key takeaways from the event? Here's a taste of what we learned at NFV Everywhere.

Moving from proprietary hardware to NFV elements is tricky
NFV fundamentally involves moving from purpose-built hardware appliances running customized software to general-purpose hardware running standardized software.

In doing so, network operators should be able to avoid vendor lock-in. In theory they should be able to mix-and-match servers and software from different vendors and play suppliers against each other for the best deals.

But proprietary appliances have their benefits, too, Heavy Reading analyst Roz Roseboro said. They're optimized for performance and hardware utilization.

Figure 2: Say Hello Heavy Reading analyst Roz Roseboro introduces panelists Rami Yaron of MEF, Red Hat's Dave Neary, AT&T Labs' Margaret Chiosi, and Intel's Frank Schapfel. Heavy Reading analyst Roz Roseboro introduces panelists Rami Yaron of MEF, Red Hat's Dave Neary, AT&T Labs' Margaret Chiosi, and Intel's Frank Schapfel.

In other words, as with most transitions, moving from a proprietary appliance model to NFV has tradeoffs.

With NFV, performance and reliability are obstacles. Overcoming those obstacles is a major challenge for NFV implementation. However, the challenge can (ultimately) be met -- and one reward is freedom from vendor lock-in.

Find out more about Network Functions Virtualization on Light Reading's NFV Channel.

Breaking vendor lock-in isn't easy
Most virtual network function (VNF) vendors have a preferred hardware partner, Nima Salehi, manager, technology strategy at Telus Corp. (NYSE: TU; Toronto: T), said. "That puts [service providers] back in the spot they were trying to move away from," because network operators are still locked in on software and hardware choices.

These hardware partnerships aren't driven by sinister motives. For optimal NFV performance, software and hardware needs to be tested together for compatibility. Also, vendors use proprietary scripts, and proprietary orchestration extensions to deliver high-performance, reliable VNFs in the absence of standards, and those extensions require hardware validation.

However, partnerships make it hard to break vendor lock-in.

Next page: Just virtualizing isn't enough

Just virtualizing isn't enough
Even after network operators have broken vendor lock-in, they still have more to do to achieve the full benefits of virtualization. Rewriting software to run on virtual machines and commodity servers just exchanges one kind of hardware for another.

That's not enough. Installing a $10,000 server to replace three $2,000 pizza boxes doesn't make sense, Masergy Communications Inc. CTO Tim Naramore said.

To unlock the power of NFV, network operators must disaggregate VNFs into component parts, Paul Boudreaux, VP Technical Strategy, Hitachi Communication Technologies America Inc. (Hitachi-CTA) , said. Network operators can deploy functions anywhere on the network, while still orchestrating them centrally for maximum operational efficiency.

The ultimate goal is to decompose VNFs into microservices, Margaret Chiosi, Distinguished Network Architect at AT&T Inc. (NYSE: T), said. "If this industry doesn't figure out how to rebuild the applications to truly take advantage of the platform, this industry will have truly failed," she said.

Containers can help fuel the journey to microservices, Chiosi said. They require fewer compute resources, making them more suitable for a disaggregated microservices architecture, she added.

There are no 'silver bullets'
An NFV service chain is only as strong as its weakest link. If a component fails, overall performance will be poor. A better hypervisor, for example, can't mitigate problems elsewhere in the chain, said Francisco-Javier Ramón Salguero, head, Network Virtualization Initiative & NFV Reference Lab, Telefónica SA (NYSE: TEF) GCTO Unit.

"In this industry, we tend to believe in silver bullets that solve all problems at once," he said. But silver bullets don't exist.

Figure 3: Lessons Learned Telefonica's Francisco-Javier Ramon Salguero describes his company's experience with NFV. Telefónica's Francisco-Javier Ramón Salguero describes his company's experience with NFV.

Even getting equipment to use the same power supplies isn't easy
Virtualization brings together IT and network operations, combining different hardware, software, and cultures. The two environments don't even share common electrical systems. Telecom uses DC power, while IT and cooling systems use AC. They need to converge, noted Geraint Davies, global director data center, Eltek.

Combining the two power systems can be difficult, but the benefit is improved scaling and flexibility, Davies said.

Next page: Culture isn't just for yogurt

Culture isn't just for yogurt
Technology issues aren't the only challenge for NFV: Organizations looking to embrace virtualization face cultural challenges too. Network and IT cultures need to converge.

SDN and NFV solve networking problems using equipment and software designed for IT. That means the teams responsible for each need to learn to collaborate.

"That's one of the biggest challenges in any organization -- how to stitch the role of CIO and CTO together, and make it a kind of CXO," Telus's Salehi said.

Crossing silos also requires developing new operational models: The Management and Orchestration (MANO) specs developed by the European Telecommunications Standards Institute (ETSI) NFV Industry Specifications Group (ISG) offer a good start towards solving that problem, Salehi said.

But the industry needs to be careful that it doesn't fragment and become unmanageable. For example, open source is key to NFV and other important New IP developments, but the open source community runs the risk of splintering and slowing the pace of network transformation: The communications networking industry needs to agree on how it wants various developments to come together in a unified platform, Chiosi said. (See AT&T's Chiosi: Unite on Open Source or Suffer.)

In addition, finding the right skill sets is a major challenge. Previously, an organization hired a network engineer or an IT engineer, and that was that, Salehi said. Now, organizations need network engineers, software engineers, programmers, scripters, cloud engineers, and people mixing those skills together. "Our engineers need to become more multidisciplinary -- probably something none of the schools can offer," Salehi said.

In the short term, companies will need to cover the "training gap" themselves; in the long term, industry, academia and the community will need to come together to combine training.

Even as network engineers become generalists, they also need to specialize. "Some still need to be laser focused on what they know. They have to know, for example, the EPC [evolved packet core] very well to virtualize it, or the whole IMS ecosystem," Salehi said.

Next page: Business models need to transform

Business models need to transform
As companies transform, customer experience needs to drive virtualization strategies. Application performance and customer needs must drive network service strategies, regardless of where technology trends are headed, said Masergy's Naramore. Virtualization is ultimately just one tool among a portfolio to serve customer needs and expectations: Virtual CPE, for example, is great for some applications, but not everything. (See NFV Everywhere? Actually, No, It's All About That App.)

And virtualization permits new pricing and usage models. For example, if an enterprise customer has a special event coming up, a network operator can spin up additional capacity on the fly, dial it down when it's no longer needed and charge accordingly, Kevin Summers, senior director, market strategy, Mitel Networks Corp. , said.

Service providers also need to tear up their old, tired telco uptime definitions. Previously, telcos looked at equipment uptime and required five-9s availability, but the metrics are different for data centers, which need to meet Tier III and Tier IV standards from the Uptime Institute.

Figure 4: New Definition Comlink's Brett Kirby says data center uptime takes on new meaning in a virtualization era. Comlink's Brett Kirby says data center uptime takes on new meaning in a virtualization era.

Now, though, the industry is transitioning to an operational (and business) model that is more forgiving of downtime, so long as the applications are running.

"We may be seeing a time when the only person who's going to care if the data center is dark is the facility manager," Bret Kirby, director of data center operations, Comlink, said. "With SDN and NFV, the application will keep cruising."

The advent of virtualization is also affecting the international balance of technology power: In addition to transforming technology and corporate culture, North American firms need to embrace the global marketplace and change their business models accordingly.

"The US is no longer the leader in innovation and change," Eltek's Davies opined. "It's starting to fall behind."

But operating as part of a global marketplace, where the potential rewards are far greater, is going to be a very tough challenge for US firms. US companies face barriers to international trade, as the US has different standards. Building regulations are different from state to state and the US uses different voltage. As a step to improve competitiveness, the US needs to adopt international standards, Davies said.

Related posts:

— Mitch Wagner, Circle me on Google+ Follow me on TwitterVisit my LinkedIn profileFollow me on Facebook, West Coast Bureau Chief, Light Reading with contributions from Elizabeth Coyne, Sarah Thomas and Carol Wilson. Got a tip about SDN or NFV? Send it to [email protected].

About the Author(s)

Mitch Wagner

Executive Editor, Light Reading

San Diego-based Mitch Wagner is many things. As well as being "our guy" on the West Coast (of the US, not Scotland, or anywhere else with indifferent meteorological conditions), he's a husband (to his wife), dissatisfied Democrat, American (so he could be President some day), nonobservant Jew, and science fiction fan. Not necessarily in that order.

He's also one half of a special duo, along with Minnie, who is the co-habitor of the West Coast Bureau and Light Reading's primary chewer of sticks, though she is not the only one on the team who regularly munches on bark.

Wagner, whose previous positions include Editor-in-Chief at Internet Evolution and Executive Editor at InformationWeek, will be responsible for tracking and reporting on developments in Silicon Valley and other US West Coast hotspots of communications technology innovation.

Beats: Software-defined networking (SDN), network functions virtualization (NFV), IP networking, and colored foods (such as 'green rice').

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like