We hope you have been enjoying the holidays and want to wish everyone a happy new year! We are excited for the opportunities 2014 will bring and we will continue to share the latest news on our blog. Our video of the week explains how the Pod Switch Interconnect is at the heart of the Plexxi pod architecture because it allows multiple switches to connect to a single pod. Dan Backman explains how these pods are strung together to form larger topologies. Here is the video of the week and a few of my reads in the Plexxi Pulse – enjoy!
TechTarget’s Shamus McGillicuddy discusses the market predictions for networking in 2014, which include many of the major trends from 2013. He says the SDN market was more hype than reality in 2013, but 2014 promises to deliver more action for SDN, NFV, OpenFlow, and others. I suspect OpenFlow will get a boost as more controllers become available. Having gear is great, but having a commercial controller product, be it proprietary or open source, makes that gear sing. I would guess OpenDaylight helps out here, even though it shows up later in the list as something to watch out for. I also agree with the comments about a dynamic network alongside overlays. I don’t think the industry can get away with an either/or approach. We will need both, which means more collaboration than we have typically seen. It needs to extend beyond supporting protocols and get more operational. Finally, I think Shamus provides an accurate analysis of the competitive landscape, especially because the battle is bigger than networking. The companies to watch in 2014 will be Oracle, IBM, and Intel.
Dan O’Shea at LightReading says the telecom industry is far from application-aware networking. Part of the reason that app-awareness isn’t everywhere is that the “Here is an API” approach basically builds a gate and asks the person on the other side to figure out how to get through it. The real work needs to be a common set of abstractions that work in vendor-agnostic environments. It is not surprising there hasn’t been more progress here because the work to-date has been either too low-level or too vendor-specific. I suspect that the new wave of commercial controllers will start to solve these problems. The endgame might still be a ways out, but at least there is movement now in these areas.