I have been researching SDN (Software Defined Networking) for a customer and then for presentation to the Cisco Mid-Atlantic User Group ( CMUG). In my opinion, SDN will be a big change for networking, similar to the change that occurred with Virtual Machines in the compute industry or virtual memory for software development.
The CMUG presentation slides and recording are available at here.
The presentation runs 80 minutes. If you’re already familiar with OpenFlow and how it works, you can skip to slide 19, where I being to talk about SDN itself. The last slide includes a set of references for further education. There are two references that I have found particularly enlightening. One is in the listed references, a presentation by Scott Shenker, “A Gentle Introduction to SDN”:
The second reference, which was published after my CMUG presentation, is from Brad Hedlund:
Scott and Brad both work for Nicira, which was recently acquired by VMware, so you may feel that the Nicira marketing message is getting too much credence. Personally, I see how the virtual networking that is described in both references can have a big impact on networking. The analogy that I used in my SDN presentation is relative to the changes that occurred when virtual memory became widespread. [Note: I describe the analogy in more detail in the recording than is evident in the slides. I’ll do a blog post about the similarities in the near future.]
There is currently a lot of hype around SDN and you can see a lot of companies trying to position themselves as an SDN player. These companies can get away with it because SDN is currently somewhat ill defined. We should see the definition of SDN become more well defined in the next year or two. You will then be able to better distinguish which companies are real players and which companies are trying to ride on the SDN bandwagon.