This list of SDx trends (software-defined “everything,” that is) to watch in 2015 is entirely subjective, meaning I’m basing it more on hope than on data. Whether these things happen, or whether they even become interesting, is up to the real world.
But from where I sit, it’s looking like 2015 will be full of fireworks for software-defined infrastructure, including SDN, NFV, open source software, and white box networking. Each of these picks might sound obvious, but hopefully the angles and motivations behind them are a little bit surprising. Feel free to leave comments to let me know what I forgot, or what I got wrong.
I love the way Peter Christy puts it in our SDN year-in-review article: The adequacy of commercial silicon could be a pivotal factor for SDN. It’s paving the way for white box switching and the Open Compute Project.
But so far, merchant chips have meant Broadcom, which enjoys a market share exceeding 90 percent. The subplot is the potential for new competition in this area. Cavium‘s acquisition of Xpliant is a factor to watch. Intel also has the means to combat Broadcom, but possibly not the motivation — its interests might lie more in the direction of its own Rack Scale Architecture.
Brite Box, Gray Box
Whatever you want to call it, the branded white box — an established vendor’s switch/router software running on off-the-shelf hardware — has arrived with Juniper’s announcement of the OCX1100. Which other vendors will follow suit? (Brocade’s original Vyatta router, which started life as open source code to run on PCs, doesn’t count.)
Deployment in 2015, Maybe
Many sources talked about 2014 being the year of proofs-of-concept, with 2015 being the year that customers — carriers, in particular — start implementing NFV and possibly SDN.
That 2014 was a big PoC year is obvious. I wonder, though, if it necessarily follows that 2015 will be the year of implementation. Certainly, some carriers are being aggressive — AT&T comes to mind, with its Domain 2.0 proclamations, and Deutsche Telekom has been talking big with its TeraStream plans. But big-picture questions such as NFV orchestration are still looming. I have to wonder if 2015 might be the year of learning-what-we-still-don’t-know, while 2016 becomes the real year of implementation.
Even More Open Source
When it comes to SDN, I do think open-source forums are replacing standards bodies. There might still be a space for formal standards, but only for dotting the I’s after the networks are already up and running. The networking industry’s embrace of open source will only get wider, or tighter, or whatever direction that metaphor is supposed to go.
By the way, have you wondered why network equipment vendors have become so gung-ho for open source software lately? I talked recently with Prayson Pate of Overture, who has an interesting take on that. I’m hoping to write it up early in January.
More startups, more acquisitions, and more Docker in demos. Docker is faster-moving and more “Wild West”-like than even SDN was. It’s being fueled by the open source nature of the code and, maybe more importantly, by the fact substantial work can be accomplished by just a few people. I’m expecting Linux containers — Docker or otherwise — to make strides in 2015 toward commercial sustainability.
A Policy-Driven World
Yes, group-based policy is at the heart of Cisco‘s Application-Centric Infrastructure (ACI). But I think there’s some good-faith work being done toward standardizing declarative networking for the good of the industry. We should start seeing signs, one way or the other, in 2015.