Virtualization of core and access networks is a key trend that will shape the communications business over the next couple of decades. In the 5G era, as core and access networks are virtualized, it will be possible for service providers, enterprises, app and platform providers to construct private networks that go far beyond today’s methods of aggregating capacity and services. And that will have direct implications for possible business models.
Consider the way mobile virtual network operators (MVNOs) now buy capacity and features, and the ways they might do so in a virtualized future. Today, the MVNO buys wholesale capacity from an underlying carrier based on usage.
But that capacity (use of cell sites, some back office support, use of voice and data bandwidth) is the same as available to customers of the underlying network provider.
The ways such capacity can be sourced will change in the coming era. At one level, the fact that all networks are virtualized means it will be possible for an MVNO or other entity to source capacity across networks about as easily as they now source capacity from a single supplier.
Google Fi (Google’s U.S. mobile service) shows, in outline, the general principle. Google Fi is set up so its user devices connect first using Wi-Fi. When Wi-Fi is not available, the customer device connects using either the Sprint network or the T-Mobile US network, depending on which network offers the “best” signal at that time, at that location.
Of course, Google Fi or other MVNOs might also, in the future, use other selection criteria, such as choosing the network that also offers the best pricing, at that moment, at that location.
Source: Nokia Bell Labs
But there is more. In the 5G era, virtualized networks will offer the ability to create virtual private networks that flow through the core network to the radio edge, and might be optimized for the particular lead application any MVNO happens to require. That “network slicing” ability means not only the ability to purchase a private network, but to specify, when required, some performance elements optimized for the MVNOs business cases.
In other words, one network might require ultra-low latency (autonomous vehicles, for example). A surgical network might require both ultra-low latency and highest-possible availability. A consumer mobile network might need highest bandwidth.
Some enterprise networks might be highly geography specific, requiring access only at some locations, and not over entire wide areas.
The big takeaway is that creating and operating wide area networks (physical or virtual) that have specific performance attributes out to the network edge (radio sites) will be possible for the first time.
New business models surely will develop, based on network slicing.
