Why software-defined networking isn't ready for the battlefield
- By Sanjay Castelino
- Sep 04, 2013
When researchers at the University of California at Berkeley and Stanford University developed the concept of software-defined networking (SDN) in 2005, it was not with the military in mind. What they ended up creating – the ability to control networks from afar, without the need for a physical switch – seems to be tailor-made for today’s modern warfighter, who is heavily dependent on IT, widely dispersed and mobile.
However, the promise of SDN may very well surpass the reality. Like many emerging technologies, SDN can be excellent for solving very specific challenges, but it may not yet be ready to solve every challenge, particularly for the military.
SDN has received much attention both in IT circles and in the technology media. In a recent report, market researcher TechNavio forecast the global SDN market would grow at a compound annual growth rate of 151.12 percent through 2016. However, the report also cautioned that there is uncertainty regarding the efficacy of the new technology, potentially posing challenges to growth.
Although SDN ideally reduces network complexity, a survey by SolarWinds indicates the opposite may be true. When asked whether SDN has the “most” or “least” impact on network complexity, more than 20 percent of respondents replied “most.”
In short, there is a great amount of interest and optimism leavened with a healthy amount of skepticism about SDN technology. All parties – including the military – should proceed with caution.
The move toward SDN is being driven by the growing popularity of the cloud and organizations like the Defense Advanced Research Projects Agency, which are investing millions of dollars in pervasive, wireless military networks. Regardless, operations are still very much driven by the need for physical technologies housed within nearby locations.
SDN is centrally managed through controllers that appear as single switches. That sounds ideal until one realizes that each physical location requires its own real switch operated by field technicians. Centralized management that could be thousands of miles away simply cannot cut it in a tactical force that depends on physical locations.
Field technicians have been heard to utter the phrase “please do not bring a 6,000 mile screwdriver to fix my network.” They’re referring to the desire of headquarters personnel to manage and fix technical operations from afar. Military personnel have been born and raised on physical technology. They’re hands-on. They like to get their hands dirty. Hence, most field operatives prefer control of their own network from nearby locations, rather than by an operator 6,000 miles away.
This is not to say that SDN is without benefits, they’re just very specific to certain situations and needs. SDN is great for highly dynamic computing environments, and excels at solving problems within fixed data centers. Indeed, a division HQ with a good size data center could benefit from SDN.
SDN also would appear to fall within DOD’s efforts to promoting a more “streamlined, rationalized and simpler network.”
However, there are still highly complex WANs that must serve hundreds of field offices and need to be managed closely. Perhaps SDN technology will mature to the point where lower level operations in each office could operate their own SDNs. We’re not there yet.
SDN still has some exciting potential, but its practical uses for today’s widespread deployments are still very much in question. And while the services are taking steps toward SDN, those steps should be taken very deliberately.
Sanjay Castelino is a vice president at SolarWinds, an IT management software provider based in Austin, Texas.