The news around NFV tends to be dominated by vendors, because there are many of them and because their product announcements represent steps toward a deployable NFV model. However, operators started NFV and they are critically important in moving NFV toward full realization of its potential.
For operators and vendors alike, the business case for network functions virtualization (NFV), meaning proving the benefits justify the investment and risk, is critical. However, there are quite a few operators that made significant advances in understanding the NFV business case, either by proving services for NFV or by building a greater understanding of how NFV services come together. Below, I've outlined six operators or groups of operators who I think have progressed the most toward that goal from the buyer side.
The NFV side
Telefónica would have to lead my list of NFV operator champions. It has been the leader in participating in NFV trials and proof of concepts (PoCs), and in ETSI NFV ISG deliberations, Telefónica has shown it has the strongest grasp of NFV as an ecosystem. Their insights into how services would be modeled and how modeling evolves into both deployment and ongoing management have been absolutely profound. Telefónica's UNICA network plan may be the proving ground for deployable NFV in 2016. Telefónica is rebidding UNICA, and the shape that the project takes will be the strongest indication so far of whether NFV can advance to broad success in 2016.
While AT&T may be best known for their role in championing the OPNFV open-source project, I think the company's strongest contribution to NFV has been an approach based on value and benefits and not on technology, even NFV technology. AT&T has worked hard to define the needs and elements of a next-gen network, and has then mapped these to available technologies, including SDN and NFV. This hard-nosed approach is paying off in a uniquely strong vision of NFV deployment.
Most network operators are operating one network, but not CableLabs. Representing the entire cable industry in the NFV process, CableLabs wields a lot of influence, and it's CableLabs representative Don Clarke who heads the ETSI NFV ISG's operator council. There's more here than political clout, though. CableLabs is taking on what might be the most important NFV target of all, residential virtual CPE (vCPE) in the form of a virtual set-top box. Residential vCPE is an exceptionally difficult model to justify, but high-value video content could be the best place to start.
Speaking of vCPE, the last of my superstar operators is Masergy, a managed services provider that launched a vCPE service in 2015. Masergy's success illustrates the fact that you can make a business case for NFV at the level of a single service, even where broad use of NFV isn't contemplated and where all NFV features aren't exercised. Masergy did select a vendor (Overture Networks) who has a complete NFV solution, and Masergy could be among the first who try to build out from an early NFV silo to a full infrastructure deployment.
The SDN side
One of the hottest technical issues in NFV is how it relates to software-defined networking (SDN), and the first of my groups of operators making a difference are conspicuous in this space. China Unicom, Deutsche Telekom, SK Telecom, Telstra and Vodafone have all sponsored important PoCs on SDN integration and it's no surprise that three of these operators are also working hard to validate NFV in the mobile space (IMS and evolved packet core [EPC]).
SDN integration could make an important contribution to the NFV process overall, because it represents an early and clear application point for "intent models," the ultimate abstraction tool that may hold the key to open NFV implementations.
NTT joins SK Telecom, Telenor and Vodafone in sponsoring key mobile/EPC PoCs. Mobile infrastructure is the largest focus of capex for operators worldwide, and the rate of infrastructure change in the mobile space is the highest. That makes mobile infrastructure an ideal candidate for modernization.
One of the key questions that these PoCs may answer is just how mobile applications of NFV would differ from cloud-based virtualization of mobile functions. Mobile elements are inherently multi-tenant, and most NFV attention has gone to single-user services like vCPE. These operators are looking at another side of NFV, a side where a lot of money will be spent in 2016.
The big NFV question
There are many positive things happening with operators in the NFV space... perhaps too many. Several operators looking to expand NFV from service-specific PoCs to broader deployment are finding that the higher-level framework of NFV still needs a lot of attention. We can deploy virtual functions, and in some cases we can prove they're manageable in single-service configurations, but the question of how to manage not one service but all possible services remains to be answered. Where will those answers come from?
That's the big question for NFV in 2016. Somehow, narrow NFV applications have to become broad deployment justifications, and there are only three ways for that to happen. First, the ETSI NFV ISG could enhance its specifications to incorporate everything needed for top-to-bottom, NFV-and-legacy, orchestration and management. Second, open-source initiatives like OPNFV or ONOS could build a complete open-source implementation, and third, operators could drive open integration models working directly with vendors.
Any of these will be a challenge to complete in 2016, yet operators are committed to try. Exactly how this next phase of NFV evolution goes will likely tell us how far NFV itself can go.
— Tom Nolle, President/Founder/Principal Analyst, CIMI Corp., special to The New IP