SDN: Waiting For The Trickle-Down Effect

The networking industry has been awash with talk of new hyper-scale and cloud-enablement technologies, such as software-defined networking (SDN). But what will move these technologies from niche solutions and cloud-scale networks to small and midsized businesses that are not necessarily focused on technology innovation? When will the small and medium enterprise (SME) market see value from SDN?

There is a well-understood model of adoption of new innovations popularized by Everett Rogers called the diffusion process. Most new technologies work their way through this process and their adoption is driven by lowered costs and increased stability and use cases. Recent examples of this model in action within the IT world include:

In this model, the early work by the innovators and early adopters sets the stage, but it isn’t until the technologies stabilize in second- or even third-generation products that they begin to gain mass-market appeal and appear less risky. For example, 10 years ago, x86 server virtualization was a relatively new market, adopted initially by very large organizations or those who needed unusual levels of flexibility. Today, just about every company that has at least two servers has virtualized them.

Likewise, several years ago, 10 gigabit Ethernet ports were expensive enough to be used only sparingly by large companies. But as prices have dropped through improved manufacturing, increased scale, and technological refinements, 10GbE has come within reach of most organizations (even if only for a few ports). This trend has brought with it improvements in virtualization density and feasibility of IP-based storage products.

In my experience of working in the SME market, these companies are generally in the late majority or even laggard stages of the IT adoption curve. Why? Simply, many companies of this size can tolerate only limited risk on infrastructure projects and are not comfortable spending time and money on a new technology until it is well-understood, well-packaged, and prices drop due to broader adoption. They just do not have the resources (budgetary or human) to be the innovators, develop new ways to use technologies, or adopt early when standards, products, and vendors are still evolving.

So what will drive SDN adoption in the SME market? I think it will be the packaging of SDN technologies into turn-key solutions. Many small companies do not have the time or desire to innovate, but they can see the value in a packaged solution that uses SDN to enable new capabilities.

The market is already starting to bear this out with products such as Big Switch Networks’ Big Tap Monitoring Fabric and Coho Data’s use of SDN to enable new IP storage capabilities in a fully-packaged product. A notable element about these products is that SDN is not a user-facing feature, but rather a means to an end.

I expect more products in the monitoring and analytics space, and possibly other datacenter services, such as server load-balancing and security solutions, to evolve using this model. An example might be an intrusion prevention system (IPS) or network traffic scrubbing platform that uses SDN techniques rather than traditional means of diverting traffic, such as policy-based routing and tunneling, to provide a more robust and flexible solution for multiple stages of traffic scrubbing or inspection.

As products like these demonstrate easy-to-implement and well-supported SDN-based systems, midsized businesses will gain comfort with the technologies, and mass adoption of SDN -- not as a raw technology, but as a means to provide new infrastructure capabilities through integrated solutions -- will begin.

I suspect this will take a few more years to further trickle down to small enterprises, but it will happen just as it has for server virtualization, high-speed Ethernet, storage area networking, and many other technologies.