With the caveats that I do not primarily cover core networks or enterprise communications, I would still argue that importance of software-defined wide area networks (SD-WANs) is not that the market is so large, comparatively speaking, or even that SD-WAN eventually will displace legacy networking platforms.
Strategically, all core networks are evolving towards virtualization, which means all core networks will define, create and support virtual private networks as a basic assumption.
In other words, all WANs eventually become virtual private networks.
There are some related advantages for service providers, ranging from the possibility of offering differentiated classes of service as a core feature of such networks, to allowing more-efficient use of networks, to reducing operating cost and capital investment.
Customers might gain from ability to buy customized network features that match user core business models (whether there are requirements for latency, quality of service or bandwidth.
In a larger sense, we move closer to the ideal next-generation network we have been talking about--and moving towards--for several decades: a network that can supply not only bandwidth but features on demand, dynamically.