Efficiencies in a provider cloud, or on customer prem (mini-cloud or edge-cloud). The operator can run a firewall, for example, on a small customer prem device that supports virtual application instances or in their own private cloud or data center environment. The choice can be made based on which is most efficient for processing power and provides the required latency.

Easier to roll out new applications. Because there is no physical infrastructure or hardware required to roll out, add or remove applications if the service provider already offers managed managed firewalls, for example, then rolling out an encryption application is simple. Just add the application to the virtual environment management tools. No additional hardware, testing, or truck rolls; only some configuration in the software tools and the application is ready to roll out to customers.

Easier to manage updates, security, patches, etc. Virtual instances are managed from a central location. If a zero-day exploit is uncovered, the operator can quickly apply security patches from the central management location without the need for truck rolls to individual locations.

Performance management or dynamic response to changing conditions. Virtualization can help by allowing the operator to provide more processing power (such as spinning up additional compute resources or additional instances of the network function application) in order to respond to dynamic load conditions. These additional computing or application resources can be “spun up” in the specific area of the network where required. Alternatively, if there are no constraining latency or traffic concerns, lesser user resources (perhaps in more remote parts of the network) can be leveraged to support higher load conditions in other parts of the network.

Eliminate vendor lock-in. Leveraging open interface standards allows customers to pursue a best-of-breed approach without vendor lock-in on the hardware or software side. With no “hardware” in place (other than servers supporting the functions), the operator can change application versions—and even vendors—as simply with one click from the management interface. No longer does an operator have to look at hardware sunk costs forcing them to continue with a specific vendor; they can change and start deploying a new vendor whenever desired. Replacement strategies, or cap-n-grown are both supported easily when the functions are virtualized (all software or licenses) and there are no underlying hardware changes to worry about.  

Orchestration, flexibility, and a look forward

In today’s world, the emphasis is on how to manage network elements more programmatically — to “orchestrate” processes that maintain, provision or troubleshoot network functionality. 

Adding NFV provides a significant amount of flexibility. There is no longer a need to deploy different physical hardware. Applications, or NFV instances, can be updated immediately through the virtual environment. However, in order to take full advantage of NFV, network operators need to focus on the management infrastructure that manages the “instances” of NFVs, in addition to the normal management and orchestration of the elements or functions themselves. Still, the upside from enabling this management and functionality provides a solid return on investment.

The evolution to a software-centric network requires a network transformation with a pace determined by operational and organizational changes rather than technological developments. An approach allowing a seamless, stepwise introduction of NFV is essential to benefit from efficiency gains in the short term while implementing the transformation steps necessary for the longer term. Open interfaces and open software are key, as they prevent vendor lock-in and cause least disruption when moving from the present to a future mode of operation.

NFV does require investment in another layer of network management: it requires you to build additional skills as you need to be able to manage the VNFs – the virtual element – in addition to managing as you do now with the functions they already provide. Even so, NFV is a meaningful way to build a truly dynamic network.

And, ultimately, when you come right down to it, that’s why NFV matters.