In my previous blog, I suggested that both Software-Defined Networks (SDN) and Network Function Virtualization (NFV), although arguably different in intent and in detail, are nevertheless both driving the evolution of network architecture in the same direction. Network functions are abstracted in order to enable them to run on commodity servers using commodity storage with all the benefits associated with that approach: cost reduction, flexibility and superior management. The "network" (the SDN data forwarding plane) will be simplified and somewhat dumb, and the intelligence, such as it is, will then exist in a consolidated management layer (the SDN control plane). From the control plane point of view, the packet-carrying network below will look pretty much like a single giant switch responding to programming and instructions from above and sending back usage events and performance parameters to the management systems. As I look at what is shaping SDN and NFV, it strikes me that we finally have the opportunity to perhaps see the reality of what could have been the manifesto for the IP revolution - it just took fifteen years to get here.

Fifteen years ago, as Internet adoption began to accelerate, David Isenberg wrote the very controversial "The Rise of the Stupid Network." He argued that the assumptions that had shaped the telecommunications industry were no longer valid: data traffic was overtaking voice, circuit switching was succumbing to packet, price performance was radically improving and customers were increasingly taking control. He suggested that the network should be "stupid," and simply deliver the bits from point A to point B and that intelligence should be the domain of intelligent endpoints interfacing on the stupid network. Today, no one can argue with the interoperability benefits of a ubiquitous protocol like IP, which has now worked itself into our smartphones, TVs, tablets, smart meters and more.

You may wonder what billing and multi-party compensation or settlement has to do with this network re-arrangement, as well as what it means to us and to others involved in billing. I would suggest that a similar seismic shift is occurring in billing and compensation as a result of SDN and NFV. In either the SDN or the NFV model, billing is still there, and in theory can be fed records of billable events in much the same way as today.

In reality, we expect the change to be much more dramatic. When an entire carrier network can be viewed and managed as a single logical switch, then clearly network functions such as usage event collection, authentication and authorization can be located in the control plane. Product managers and network managers will be able to implement services by programming the network, following a testing and validation regime no different from that used for implementing software releases. In other words, it will allow for faster, less risky service development and implementation.

It is conceivable and, in my opinion, very likely that product managers, liberated from the tyranny of service silos and instantiating new services across silos and their associated network elements, will quickly take advantage of this power and flexibility to dream up all sorts of new network services and service mashups. With these developments, they will likely bring in non-network service components, and with those services, a previously unimagined array of novel chargeable events.

How will billing providers react to this newfound flexibility? Or more specifically, how will traditional billing hold it back? Find out in Part 4.

• SDN: Evolution to Special Purpose Clouds Challenges Traditional OSS/BSS
• Close Familial Ties: SDN and NFV | Part 2
• SDN and NFV: End-User Service Benefits Is What Matters | Part 4