Software-defined networks (SDNs) are still in the honeymoon phase, seen as the cure-all that will solve many, if not all, of the capacity and efficiency problems of enterprises and service providers. But the difficulties and limitations of designing and deploying SDNs are not yet apparent.
Governmental networks have long been a proving ground for telecommunications and IT technology. Indeed, the Internet itself is largely the product of federal and academic research and testing. Some of the smartest people work for the government. In addition, the profit and loss equations that control research and development and infrastructure investments in the private sector are different, and significantly loosened, when Uncle Sam is footing the bill.
The problem is that the government wants to use SDNs but, at the same time, the people who are running things are fairly ignorant of the SDN technology. That was the dual finding of a study conducted by Wakefield Research for Juniper Networks. The interest is real: Thirty-four percent of respondents plan to adopt SDN within two years, 91 percent see at least one place in their organization that SDNs would be beneficial, and 61 said it is an important factor in network buying decisions.
So the desire is there. The problem is that the knowledge isn’t, according to the eWeek report on the research:
Sixty-one percent of the 250 government IT professionals surveyed by Wakefield Research said they are unfamiliar with SDN, and even among those who are familiar, 58 percent had two or more misconceptions about the technology. For example, 11 percent said SDN applied only to data centers (it doesn't) while 15 percent believed it required manual configuration of hardware (it actually is designed to remove manual configuration of the hardware, which will make the networks more dynamic, flexible and agile).
Pieces at Forbes and CIO confirm the hypothesis of the Juniper/Wakefield research, highlighting how diffuse and confusing the nascent category is.
Forbes’ commentator Patrick Moorhead describes the SDN landscape as one that is still forming and one in which vendors have decidedly different strategies. He outlines the strategies of Cisco, Hewlett-Packard, Alcatel-Lucent, Dell and Juniper and links to a longer story at his website that dives deeply into the state of the sector.
Another sign of the immaturity of the market is the number of startups that are vying for support. These companies want to gain traction and push the technology in their direction. At that point, they either would become long-term players or sell out to the bigger networking vendors and have a huge payday. The companies whose approaches are profiled in the CIO story are Big Switch Networks, Embrane, Jeda Networks, Metacloud, Midojura, Pertino, Plexxi, PLUMgrid and the OpenDaylight Project.
SDNs are an important next step for networking that many executives want to take. At this point, however, the precise direction of that step is still unclear.