SDN has been talked about for some time now and we’ve seen it go from conceptual to production. In 2014 we saw a lot of Data Center SDN, but what I’m seeing now is that there’s a focus on the WAN happening. Why? How?
Why A Focus On WAN SDN?
Let’s start by talking about the dollars. Many IT organizations spend a great deal of money on WAN bandwidth. A WAN circuit with lower bandwidth that what a residential user can get may cost $2000.00 plus (easy) per month. Factor in multiple sites, various SLA requirements and so forth, and you start to see the pocket book take a hit. The exorbitant price paid for this bandwidth could undoubtably find better use elsewhere in the organization. So the resoning behind SDN in the WAN is apparent from one angle; Lower the OpEx!
How Is It Achieved?
Consider the way we do things now. Today purchase a WAN circuit with an SLA and guarantees on bandwidth, throw a ton of money at it when we need more bandwidth and life goes on. With SDN we can abstract ourselves from the WAN provider by using overlays. Additionally with SDN we can achieve full utilization of the WAN links that normally sit and pile up dust.
Who’s Doing it?
So who’s positioned to be a player in this field? Cloudgenix is one such company offering a SDN solution for the WAN, and they’re scheduled to speak to a number of Networking Field Day 9(NDF9) delegates in February. Velocloud is another who plans to reinvent the WAN, and they are also slated for NFD9.
I think there’s real value in the SDN in the WAN, especially in the way of analytics. Application level visibility is something that you don’t get in current WANs and that you totally lose with IPSec VPNs. We’re not going to see this take off like a wild fire this year, but it’s going to begin to emerge and likely will become more prevalent in the next 5 years. Provided the cost of thew solution is inline with a reduction in OpEx and does indeed provide a scaleable solution with the visibility we require in todays application driven environments. Integration with popular cloud providers would also make the solution very appealing to some. I’ll know more after NFD9, but I think the solution could be viable if implemented properly.