Contributed article by Sunay Tripathi, Co-Founder and CTO, Pluribus Networks

2014: Vendor Lock-Ins on Data Center Top-of-Rack Access Will Begin To Crumble

As most of you have already noticed, these are exciting times for the networking industry. In this article, I will look at Data Center Top-of-Rack (ToR) Switching, driven by the gradual adoption of 10GbE in the server access, which, I believe, has been the most dynamic and innovative segment of the whole Ethernet Switching market over the past five years. Despite falling port prices  (-75 percent in five years) and a dramatic rise in port volumes (~30X over the same period of time), this segment continues to be dominated by traditional vendors whose business model is predicated on selling proprietary hardware platforms at a hefty profit margin.

Based on what I am hearing from our customers and partners today, I believe that in 2014, we'll finally see signs that the inescapable commoditization of Data Center Top-of-Rack switches will begin to accelerate, and this segment of networking will finally be ready to follow in the footsteps of the server market. 

I don't want this to be another "empty" prediction, so let's discuss what needs to happen for this prediction to materialize in 2014.

I think that 2014 will mark the inflection point where merchant silicon, rallying behind the release of Broadcom's Trident 2, will break away in 10GbE port market share. Traditional networking players have not been able to keep up with the torrid pace of innovation set by merchant vendors over the past few years and all of them, without exceptions, are now forced to embrace the disruptive economics of merchant silicon. The availability of powerful merchant silicon has also lowered dramatically the barriers to enter the 10GbE ToR hardware business. Over the past few years, an increasing number of white label vendors have been offering low-cost ToR switches.

The key question is: what has prevented merchant silicon-based white boxes to go mainstream despite radically better market economics?  The answer is software. Enterprise-class network operating systems are the single biggest advantage of the incumbents and the single biggest weakness for white boxes. While the incumbents have perfected over the years their network operating systems, the software on white boxes continues to lag behind in quality and L2/L3 functionalities.

I expect 2014 to mark the beginning of a sea change in network operating systems for white-label hardware. Enterprise-quality, feature-rich network operating systems will emerge. Business models will evolve towards software monetization, the high profit margins commanded by the incumbents will erode, and, finally, the ToR hardware market will march down the same commoditization path of the computing a decade ago.

##

About the Author