Market oracles expected the relatively rapid adoption and growth of NFV by service providers to be led by the increasing capabilities of NFV management and organization (MANO) solutions along with the support of open source organizations that would help propel technological advances. This would ultimately lead to customer success stories, either with full-blown proofs of concept (PoCs) or deployments.
Of course, all predictions entail a certain number of uncertainties. So let’s take a look to see how well we did.
NFV Service Provider Market Growth: Did It Happen?
The global carrier SDN and NFV hardware and software market was worth less than $500 million in 2013, according to IHS technology. IHS now forecasts the global network functions virtualization (NFV) hardware, software, and services market to reach $11.6 billion in 2019, up from a very healthy $2.3 billion in 2015. But it also points out that the service provider space is just beginning a 10- to 15-year migration to virtualized networks.
Michael Howard, an analyst with IHS, pointed out at the SDN World Congress that much of the revenue moving into SDN and NFV would be “displaced revenue” – capital spending that is being moved from one area to another. Howard projected $30 billion of service provider capital spending on software-defined networking (SDN) by 2019, with only $6 billion of the total representing new capex.
“Sixty percent of it is what they would spend anyway,” said Howard. “There is displaced revenue.”
It’s clear that the interest, energy, and speed of this market is real. But real revenue-generating solutions and cost-saving infrastructures involves a slow transformation process that will have to build in the years to come.
Did Developments in NFV MANO Provide a Foundation for Adoption?
Key milestones for NFV MANO’s maturation in 2015 included the GitHUB project OpenMANO, led by Telefonica, and the Open Platform for NFV (OPNFV) expanding its original scope to include this critical solution area. Additionally, at Mobile World Congress 2015, multiple vendors highlighted their partial or end-to-end MANO solutions. Approaches to organizing NFV orchestrators, VNF managers, and virtualized infrastructure managers (VIMs) varied considerably, which raised questions about how various CSPs’ and vendors’ MANO architectures would work together in the future.
New technologies pose risk, which to service providers is a four-letter word. It became clear that in order for NFV deployments to be successful, NFV MANO is indeed a key component but not the only one. NFV also needs to integrate and be reliable and secure. Other components such as OSS/BSS systems must be more tightly integrated to make a complete solution, and because of that, actual deployments have been more scarce than anticipated.
Ultimately, 2015 turned out to be the year of NFV technologically “storming” – and we believe we will see NFV “forming” very soon.
Did the Open Source Organizations Help or Hurt?
Today 38 multicompany, public, ETSI PoCs have been initiated or completed. (Check out the SDxcentral Open Source Projects Directory to see what’s going on.) They target common NFV framework architectural components that the industry has agreed on under the auspices of the ETSI NFV Industry Specification Group.
It was evident at the OPNFV Summit 2015 that virtual network functions (VNFs) that are being deployed first are vCPE (in both residential and enterprise), vEPC, and vIMS, particularly for VoLTE usage models. These offers are specific, lower cost, and have limited risk compared to entire infrastructure transformations.
There is so much interest in the NFV market that many different vendors are scrambling to provide components for NFV infrastructure, which is adding to general complexity. If we are going to achieve the ultimate industry goal of an open, standards-based solution, we need to ensure interoperability across, not only new solutions being developed, but also between new and standing network infrastructures. This is where the open source community has made great strides with the implementation of labs, such as the OPNFV labs, which are critical venues for technical collaboration.
These will become increasingly important as we proceed into 2016.
Who Deployed Services With NFV in 2015?
In the United States, AT&T in 2015 was a prominent NFV developer with its Network on Demand project, which uses NFV to create faster on-demand services. It’s working with Brocade, Cisco, and Juniper Networks. In addition to focusing on what has already been labeled as one of the first deployable SDN/NFV solutions, vCPE, AT&T is working on the lower part of the network stack by “pushing NFV down the stack.” It hopes for lowering costs in the future.
It was not all successes, however, as we saw at Telefónica. Telefónica has dropped Hewlett Packard Enterprise (HPE) as the technology lead for the ambitious Unica project, a network makeover centered on NFV. Telefónica hasn’t elaborate on its reasons, but it’s worth noting that the Spanish carrier was one of the few big names that didn’t join the OPNFV open source initiative.
Even early in 2015, Alcatel-Lucent spoke about roughly 30 engagements for its CloudBand NFV platform. Equally clear was that only about six were in production. The company was also not immediately ready to discuss these in much detail at the time.
What Did We Get out of 2015?
As an industry, NFV, along with SDN and their tightly linked relationship, grew rapidly in many areas including the number of industry vendors participating in development, open source organizations, and some actual deployments of NFV solutions. Key industry events were held, and ideas were shared – such as which solutions are viable today like vCPE and vEPC. And understanding the complexity and security issues became clear while work has begun.
Perhaps we’ll look back and refer to 2015 as the year we came further than we thought about as an industry, but did not meet the revenue generation and cost savings we had originally anticipated.
2015 was the NFV Year of the Storm.