By Jim Theodoras
OFC/NFOEC 2013 is rapidly approaching, and those with young kids (or an inner child) are surely looking forward to a return to Anaheim due to a nearby attraction with a large talking rodent as its mascot.
Talk of Software Defined Networking (SDN) will be everywhere as the networking revolution begins to transition from concept to reality. Make no mistake, SDN is a revolution that will both kill old business models as well as create new ones. Unfortunately, it is also incredibly uninspiring to behold, which is kind of the point. The entire network-- first Ethernet switching and now inclusive of optical transport-- is abstracted to a simple connectivity table that an external controller can act upon. This is great if you are a service provider looking to improve your network, but not so great for entertainment value. At least there will be free swag in those booths.
100GE redux will also be on display as the industry feverishly works toward a more palatable client solution for end customers. While the C form-factor pluggable (CFP) Multi-Source Agreement (MSA) was great as a proof of concept and makes a nice platform for power hungry line side transceivers, it has arguably failed the market test as a client port, hence the development of CFP2/4. Meanwhile, the Quad Small Form-factor Pluggable (QSFP) package had it right from the start and has demonstrated steady evolution from 10, 40, 100 and now talk of 400GE, with wide adoption from telecom, datacom, storage and HPC markets with support for Ethernet, Fibre Channel and InfiniBand protocols—sort of the whole reason an MSA exists in the first place.
Speaking of 400GE, the first vestiges of 400GE will begin to appear. 400GE efforts appear to be leveraging technologies being developed for the 100GE do-over, with faster, narrower electrical bus widths and smaller pluggable packages. While I fully support 400GE efforts, I do in the back of my mind wonder if we are at a historical point of transition where Ethernet joins other protocols like Fibre Channel and Optical Transport Network (OTN) by growing in smaller multiples than 10, i.e. 2 or 4 respectively. In a way it makes sense, as chips speak binary and bus widths also tend to be in multiples of 4. The smaller step makes the transition easier, but smaller steps also mean a vendor may have a harder time differentiating themselves in the market. While everyone on the show floor will have 400GE on their mind, perhaps we should give some thought to what comes after it.
It is dangerous to opine that we are approaching an asymptote in Ethernet speed, as all previous predictions quickly fell victim to the endless enthusiasm and brilliance of engineers worldwide. How many people thought long-haul 100G transmission would never happen? Today it is considered routine. Yet, barring the next unforeseen leap in technology, it does appear future increases in electronic bus speeds, laser emitters and fiber capacity will be more challenging, even with the help of Moore’s law.
Perhaps this is where optical waveguides on printed substrates take over. Or perhaps silicon photonics actually reach their as yet unrealized potential and save the day. Or maybe an unheard of technology springs from the shadows to keep fiber optic communications on its blistering pace of growth. If and when that does happen, there is only one place to see it first… at OFC/NFOEC.
Jim Theodoras is senior director of technical marketing at ADVA Optical Networking.
Posted: 19 February 2013 by
Jim Theodoras
| with 0 comments