Hard Truth #1: Service providers expect that virtualization software will be inexpensive or free. While they won’t achieve that goal in many cases, they will require that the software solutions cost less than today’s appliance-based solutions.

Hard Truth #2: How we license NFV software must be flexible enough to address varying business models, sometimes residing within the same organization. Older licensing models such as node locking won’t work.

Hard Truth #3: Service providers want NFV, but they will wait to get the licensing arrangement they want.

What does that mean for suppliers of software Virtual Network Functions (VNFs)?

The Future of NFV is About Innovation, but the Present is About Savings
One of the most important aspects of NFV is the splitting of closed appliances into separate hardware and software components, enabling independent selection of each. In addition, the software VNFs are designed to be assembled into services and controlled via automation, enabling rapid innovation. That’s the future benefit of NFV. But, what about today?

Service providers expect that NFV will immediately lower CapEx and OpEx. That means that the cost of a server and VNFs should be less than the cost of a set of appliances, with savings of 25% or more. Otherwise, the cost of integrating and deploying a virtualized solution is too high to justify the theoretical future benefits.

During the trial and proof-of-concept phase with NFV and the focus has been on, “Will it work?” As our customers move to deployment, their focus changes to, “Does it cost in?” We know from experience that service providers will wait for the right answer to that question … for a long time.

Lower Starting Cost is Just the Beginning
In addition to a lower cost for a virtualized solution, service providers expect that moving to a software-based platform will give them a variety of other cost advantages:

• Different VNFs for different applications. Service providers may be willing to pay for a premium VNF in some cases but not others. For example, they may be willing to pay licensing fees for a full-featured firewall/security application with email and URL filtering. However, they may want a bare-bones open source solution for low-end applications. Big idea #1 for VNF suppliers: Getting some of the business may not mean getting all of the business.
• Usage-based licensing. End users want to pay as little as possible, and preferably want services for free. When they do pay, they expect to pay only for what they are using, and only when they are using it. Big idea #2 for VNF suppliers: Create licensing models with fees that vary with usage, whether in terms of features, bandwidth and/or time, and minimize the upfront payment.
• Shared risk/shared reward. One of the great attractions of a virtualized service is that the incremental costs of enabling a service can be very low. This fact enables service providers to emulate cloud-based service and offer free trial periods. However, they don’t want to pay for licensing when they are not getting any revenue. Really big idea #3 for VNF suppliers: Accept contracts where you don’t get paid until billing starts for the service, and then take payments based on a share of the revenue.

Nothing is Easy
These hard truths and big ideas will cause headaches and heartaches for VNF suppliers. However, the pain is just starting. The following complexities of licensing must also be addressed:

• Integration with orchestration and billing systems. Elastic services require scalable VNFs, and the scaling occurs under the control of an orchestrator. The orchestrator may need to tie into license management to validate the scaling event versus the licensed features and capacity. Likewise, the billing system must be informed of any changes in a usage-based service.
• IP protection. Licensing solutions must meet the new service provider requirements while not being onerous to implement and also preventing unauthorized use of the VNFs.
• Revenue recognition. VNF suppliers must connect their license enablement solutions to business or CRM systems to properly recognize what and when to bill customers, and how much can be recognized as revenue.
• Difficult to size the market. We now know how big the markets are for the various appliances being used to deliver services. When they are replaced by virtualized solutions, how much will the revenue per usage go down? How much will the total addressable market go up? These questions are hard to answer, and as such make it difficult for VNF suppliers to plan their business.

Face it: To Accelerate NFV Deployments, Licensing Models Must Change
The hard truth is that we VNF suppliers can bring tremendous innovation and growth to the telecom industry. Doing so requires accepting that licensing as we know it needs to become licensing as the service providers want it. We’ll be watching innovative suppliers take the lead and others taking their leave.