New, disruptive technologies generally are incubated best in small, entrepreneurial start-up companies. However, selling these sorts of technologies can become a daunting task. One example of a highly disruptive technology is software-defined networking (SDN).
SDN has been a glimmer in technologists eyes since its original inception shortly after Sun Microsystems launched Java in 1995. But it wasn’t until 2011 that the Open Network Foundation was created to promote SDN technologies. And, today in early 2015, the technology still hasn’t quite taken off in a big way. However, most technologists and analysts believe it will be at least another year before SDN deployments become more common place and that it will be another 3-5 years before the market matures.
The advantages of SDN are quite substantial over today’s technologies. With current network architectures, applications must conform to the rigidity of the network. As billions upon billions of new connected devices are added, current technologies will be unable to manage this new environment. Signs of this are already happening. With SDN, its 180 degrees opposite, where the networks becomes a natural extension of the application. Eric Schmidt, recently commented that “the internet will disappear,” which in essence is saying the same thing. Connectivity will just become a natural state.
Fundamentally, SDN, especially overlay technologies are much simpler, more scalable, and more secure than today’s equivalent technologies. Connectivity becomes embedded as part of the application. SDN provides a path for companies to dramatically reduce network costs and become more agile and innovative.
So if the advantages are so great, why is it taking so long to see adoption?
I’ve identified at least four risks: technology, market, operational and vendor risks.
SDN is new technology. It is not well-understood by CIOs and the IT management teams. They need a tremendous amount of education just to understand how to operate an SDN environment. In addition, it hasn’t been fully proven and field tested so CIOs are rightly concerned in wanting to know it works as advertised. Since the technology is an inline solution they can’t afford to add new technology that may bring down their communication networks or create new security threats.
Risk averse buyers, such as CIOs, like to buy when other similar buyers that are like them have already bought and successfully implemented a similar solution. Most don’t want to be first to market and want to see others work out the kinks of a new markets and new technology. These buyers like to do a lot of research and talk amongst their own colleagues at other companies to learn what challenges there are, what works and doesn’t work, what things they need to consider, and which vendors are best to work with. This is a very slow process.
While current network technology has been around for 10-20 years and is ripe for change, it’s also often easier to stick with something you already know and understand than to jump to newer technology. Perhaps even more relevant is that Cisco and other established players provide FUD in the market since there is a lot of potential downside impact to their businesses moving too quickly.
Most CIO’s operate to an annual budget. They know how the business is expected to grow and what their needs will be over the course of the year. It’s relatively easy for them to anticipate how many more servers, routers, switches, load balancers, and the like to buy, however, because SDN is so new, they don’t yet know what it will cost them to deploy.
Even if they fundamentally believe in the benefits of SDN, they don’t yet understand the investment costs to achieve those benefits. Supplier pricing models are different. Also a challenge to vendors is if IT hasn’t calendared a project for the budget year, you are fighting other established priorities for the business.
And because SDN changes the very nature of the role of network management and cyber security incident response teams, they may not have the proper skill sets in place. Cisco has done a great job indoctrinating CCIE’s in a way to think about networks, its going to take them time to understand a software-based approach when they are used to lots of hardware boxes. So, not only does SDN introduce new products into the environment, it may require substantial changes is processes and training.
Most new SDN technologies have been launched by small start-ups. Who’s going to win? Which ones can CIOs depend upon to survive when they are about to make potentially dramatic changes to their networks.
What does this all mean? Some disruptive technologies can cause change quickly. The calculator replaced the slide rule rather quickly but that still took several years to accomplish. Disruptive technologies that have broader implications across a company will take much longer. This is why most analysts predict SDN to take another 1-2 years for live, broad based deployments and another 3-5 years to begin to reach maturity.