SDN Still Has to Prove Itself to the Enterprise
SDN must overcome several key challenges before it can make it through the hype cycle and establish itself on the other side.
The software defined network (SDN) is the newest member of the virtual enterprise family, and as such has been the recipient of a lot of oohs and aahs and predictions of how big and strong (and smart) it will be someday.
But first it must endure a few growing pains. One of the more difficult will be gaining the respect, or at least the understanding, of its elders. Even among highly experienced IT professionals, it seems SDN is still something of an oddball.
According to Network Instruments, SDN has been the most talked-about development for three years running in its annual survey of networking professionals, although consensus as to what it is and what it can do for the enterprise is clearly lacking. In the latest breakdown, 37 percent of respondents labeled SDN as "undefined, like a trip without a road map," while the next largest portion, 34 percent, cited automated provisioning as its defining characteristic. 20 percent are planning their initial deployments within the year, even as skepticism and a desire to "ride out the hype" hold sway over a large plurality.
In the short term, this could be a problem for the networking industry: the more severe and longer-lasting the FUD, the more it hurts the bottom line. This is due to two factors. The first is that enterprise executives are loath to trust valuable data to unproven technology. The second is the disinclination to flesh out existing infrastructure if something brand-new is right around the corner. Infonetics Research calls the latter "SDN hesitation" and claims it can already be seen in service provider and carrier markets, where deployments of edge, core and long-haul Ethernet switches were down 13 percent in the first quarter.
Indeed, it seems many vendors are way out in front of SDN technology while customers are still trying puzzle out what it all means, says VMware’s Martin Casado. The first thing to keep in mind is that SDN is not merely a technology upgrade but a fundamental shift in the way data and data infrastructure is developed and handled. In that vein, he said, it would be wiser for the vendor community to focus more on making SDN a usable, viable alternative to current networking technologies rather than blazing new trails in functionality.
A key problem for SDN is that fact that network professionals are starting to wise up to the fact that its key attribute, separation of the control and data planes, is really not that much of a game-changer, says F5 Network’s Lori MacVittie. Sure, you can move much of the provisioning and configuration to the control place, but without a sophisticated management and automation stack, all you’ve done is shift complexity from one piece to another. The challenge going forward is not simply to implement SDN but to recognize the fundamental ways in which it can open up entirely new ways of doing things and then tailor your automation and orchestration capabilities to suit those ends. Remember, if SDN does not lead to a noticeable change in application performance, then something is wrong.
None of this is to suggest that SDN produces challenges any greater or less than other paradigm-shifting developments. After all, how many PCs sat in the corner of the living room collecting dust during the 1980s after users realized they didn’t really know what they wanted to do with them?
But it does mean that SDN still has a way to go before it can claim to be a truly revolutionary development. And the final judgment will rest not on what it is, but what it does.
Photo courtesy of Shutterstock.
Arthur Cole covers networking and the data center for IT Business Edge. He has served as editor of numerous publications covering everything from audio/video production and distribution, multimedia and the Internet to video gaming.