This week, the Optical Fiber Conference (OFC) took place in Los Angeles. OFC always brings back memories of back in the days of the telecom bubble, like OFC 2000 in Baltimore when you could show up with a suitcase full of MEMs (Microelectromechanical systems) and get a billion-dollar valuation.
This leads me to the next question: Now that software-defined networking (SDN) is the next big thing, will it be as big as we imagined optical networking would be in 2000? Is it another bubble? Will it crash like optical? Or is it a whole different thing — maybe a world in which optical networking and SDN can create the next big thing together!
As I’ve maintained in my analysis, SDN is the biggest paradigm shift in networking in decades. I admit, as soon as you use the phrase “paradigm shift” you see customers rolling their eyes and investors running for the exits. But I don’t think it’s overhyped in the way optical was, and I do think this revolution is different — it’s more rationalized and realistic.
Today let’s run through how SDN is not like the optical telecom bubble of 2000, and how SDN might actually merge with optical to build the more automated and intelligent network we were looking for in the first place.
SDN: Show Me the Money!
The excitement around SDN is more subdued and realistic than in the optical bubble days of 1999-2000. We don’t see George Gilder moving stock prices of Chinese component producers by 20% or non-revenue companies going public with multi-billion dollar valuations. What we do see is a lot of companies carefully spending money, working on standards, and developing new software that produces demos and interoperability work.
We all know that the customers, enterprises and service providers, aren’t exactly whipping out their checkbooks for billion-dollar multi-year RFPs, like a Bernie Ebbers-run Worldcom world. The spending is more cautious and pegged to important landmarks.
The SDN startup world, therefore, is very different than it was in 2000. Those working in startup land will tell you it’s a grind just like any other job, working long hours to get a technology to the point where customers and investors can see some value. Startup valuations are in the low-hundred millions in general, not the billions.
This is good. This means the industry is being built brick-by-brick, for rationalized business cases.
More importantly, the SDN story is about the optimization and automation of networks. This is a stronger Return on Investment (ROI) story than what we heard during the optical bubble days, which was more about building out speculative capacity that may or not have ever been needed (Global Crossing, anyone?). The networking landscape now is about finding more efficient ways to move the bits around and gain business advantage through the automation of networks, not creating bandwidth for bandwidth sake.
The companies that do well in SDN, I believe, will be the ones that have the strongest migration story, rather than trying to sell you a bright, new shiny piece of software. The scary thing about this is that the largest companies know this, and they are looking at SDN as a way to actually reduce spending, rather than increase. So if you are selling SDN software you have to be very careful about what the end result is, because it may cannibalize other businesses.
Coming Soon: Optical SDN Integration
The utopian vision of SDN is that somehow it merges with the futuristic vision that was optical network back when we were all younger and greener. Optical network was about moving bits at the lowest possible cost, and the ease of manipulating wavelengths with Dense Wave Divions Multiplexing (DWDM).
We’ve seen big gains there. If you were to jump from OFC 2000 Baltimore to OFC 2015 LA, you would look at the current products by optical players such as Ciena, Cisco, and Cyan as natural progressions to the future. Today they’re now talking about 100G metro systems not the 10G metro systems from the year 2000. And we are talking about optical-packet integration on a level that did not exist in 2000.
The worlds of traditional IP networking, SDN, and optical are integrating, albeit slowly and methodically. And you will see more of this over time.
Yesterday I spoke with Gerald Wesel, the CEO of Polatis, an optical networking switch company that came to life in 2000, survived the crash in the mid-2000s, and is now finding a market for its optical switching in the data center. He says it’s been an interesting ride, one in which the optical markets, boosted by data-center demand for intelligent and scaleable bandwidth, are now starting to rev up again.
“Optical networking never really went away, but it went quiet for a while and now it’s resurging with video as the key driver. It’s about moving this around the metro, data center optimization, and SDN. Without SDN, optical networking is just a patch-panel replacement.”
That’s right, to move to the next level, optical can’t be just about switching wavelengths in a futuristic patch panel. Just having flashy all-optical networking technology in itself. It’s about applying the automation and intelligence of SDN to build a smarter data center.