In consumer marketing we are familiar with the concept of giving a way one product (or selling it at a heavily discounted price) in order to build a market for even larger sales of another product. Giving away the razor to sell the blades is the classic example, and a more modern one is giving away the printer to sell the ink cartridges.
In telecom, the sales model is quite different. Every product gets sold on its own merits, and its own business case. If anything, a reverse process takes place, where value-added features are given away or sold at a deep discount as a sweetener for the main product. For example, we can sell a high speed coherent transceiver, which comes with an integrated performance monitoring option, that the customer may demand for free rather than paying for.
Selling an SDN control and applications infrastructure (for the WAN), however, may justify using a razors to blades model.
SDN’s problem is that it is still primarily regarded as technology that exercises centralized real-time control over specific network domains. There is recognition that this centralized real-time control can be enabler for value-added applications like multilayer provisioning, restoration, and optimization, or specific services like bandwidth-on-demand or a more generalized Network-as-a-Service (NaaS). However, the obstacle to overcome, in order to start building a business case based on these applications, is that a critical mass of SDN-controllable networks elements must be in place to control. Only then can you start justifying an SDN control layer and associated applications.
This creates a dilemma. Service providers may be reluctant to deploy SDN controllers and application orchestrators until a critical mass of SDN NEs are in place. In the same vein, SPs may question why they should specify that NEs are SDN-ready unless there is a control infrastructure for them to home into.
The way out of this vicious cycle may be to duplicate the razor-blades or printer-cartridges model. Vendors should give away the SDN Controllers and baseline applications, along with the starter kit of SDN-readiness interfaces on any NEs that are sold until a critical mass of value is reached.
Then and only then, once a minimum baseline of value is recognized, can we actually start selling SDN capabilities – additional control infrastructure, suites of SDN applications, and SDN controllable features and interfaces on the NEs. Of course, there will be pushback (in particular, and perhaps rightfully so, for the basic SDN interfaces) but at least there is now a real discussion around realizable value and near-term return on investment.
Giving away SDN “starter kits” may be the only practical way to prime the pump for selling SDN infrastructure and applications in the WAN.