We normally think about network functions virtualizati0n as a tool network operators can use to create new capabilities while lowering cost. What we tend not to think about is whether NFV will create new ways for big new competitors to enter mobile service provider markets.
NFV could allow big and powerful third parties to federate multiple networks, creating their own virtual networks out of capacity sourced from multiple physical networks. It is not clear that doing so, at scale, would necessarily provide an operating cost model that is superior to a facilities-based approach.
But doing so could provide enough value that the ultimate business model is superior to that of owning and operating facilities.
Source: Nokia Bell Labs
If you think about it, even without using NFV, Google Fi combines access assets of Sprint, T-Mobile US and any local Wi-Fi as the foundation of its mobile phone service.
Google Fi already federates Wi-Fi, Sprint network and T-Mobile US network access assets in support of its mobile phone service.
Ultimately, what matters is whether the additional value of federating assets is high enough to outweigh the costs of leasing access on the underlying mobile networks, in relation to the business value contributed by the actual phone service.
It is commonplace to hear arguments that owner’s economics eventually become necessary for any leading mobile service provider. Whether or not that is true for all future leading mobile providers is an open question.
So far, virtualization arguably has posed a threat mostly to suppliers of network infrastructure systems, software and hardware, as virtualization allows service providers to avoid buying more-expensive branded gear and rely instead on open source and white box solutions, while also reducing the amount of intelligent network elements scattered around the network.
By separating the data plane from the control plane, networks can centralize control functions while shifting to dumber appliances in the field, reducing overall cost.
But we should eventually see something possibly unexpected, when many major networks are virtualized, and that is the ability of any entity to construct a big virtualized and custom network of its own.
While it is true that any such effort would not have owner’s economics, that might not be so important for an entity with a different business model, not reliant on access revenues, but for which the ability to integrate access with the other business functions creates more total value.
In the past, large networks with scale have found ownership more affordable than renting. In the future, that could change, but possibly mostly for third parties with business models based on content, transactions, advertising or app and service products beyond network access revenue.
So perhaps NFV, though a tool to increase agility and reduce cost for any single service provider, might also serve in the future as a way for third parties to create communication services in a different way, much as IP messaging, video communications and voice are features of apps that create revenue other ways.