Planning for 100G & Beyond
At OFC/NFOEC last week, I had the opportunity to speak on an Optical Internetworking Forum (OIF) -sponsored panel entitled “400G vs. 1 Terabit: Market Needs and Technical Challenges.” The session focused on the next-generation transport speed beyond 100Gbit/s.
When invited to speak, my initial reaction to the panel was: Isn’t this way too early? To put things in context, 100Gbit/s transport has yet to even get off the ground. Heavy Reading has just wrapped up several months of research on the emerging opportunity for 100Gbit/s – described in our latest report, "100Gbit/s Transport: Forecast & Analysis." The 100Gbit/s Ethernet interface was standardized by the Institute of Electrical and Electronics Engineers Inc. (IEEE) in mid 2010, and there were only a handful of commercial deployments worldwide throughout 2010 – all involving just two equipment suppliers, Alcatel-Lucent (NYSE: ALU) and Ciena Corp. (Nasdaq: CIEN). Our forecast calls for a 100Gbit/s ramp-up by 2012, with the transport rate exceeding both 40Gbit/s and 10Gbit/s in backbone capacity by 2015. We expect 100Gbit/s to have a long future as the dominant transport rate for backbone networks.
So, why are we talking about 1Tbit/s? While it’s absurd to think about a near-term commercial market for 400Gbit/s or 1Tbit/s transport, from a research and development perspective, the time to get started on the next-generation transport rate is now.
First, a fundamental issue needs be resolved as we look beyond 100Gbit/s: What will that next bit rate be? This is a new challenge for transport. Historically, telecom has migrated in 4x increments and datacom (Ethernet) has migrated in 10X increments. While 100Gbit/s Ethernet was a highly successful collaboration between the International Telecommunication Union, Standardization Sector (ITU-T) and the IEEE, the trajectory forward needs to be determined – hence, the 400Gbit/s (4x) vs. 1Tbit/s (10x) discussion. Collaboration between the IEEE and the ITU-T, and perhaps even the OIF, will be required.
Second, technology challenges continue to mount as transport rates move higher. The game of DWDM transport changed dramatically when transport moved beyond 10Gbit/s to 40/100Gbit/s, because of fundamental physics challenges that became significant beyond the 10Gbit/s bit rate. To overcome these challenges, we’ve seen tremendous innovation in modulation formats, the use of coherent detection and advanced forward error correction. Beyond 100Gbit/s, maintaining performance becomes even more difficult, while yielding more capacity from a fiber strikes a path of diminishing returns (regardless of performance). The best solution is yet to be determined. There are also challenges on the client side: Groundbreaking innovation will be required to reduce form factors, power consumption and costs.
Third, the pitfalls of 40Gbit/s need to be avoided. A major problem with 40Gbit/s – one that persists – is that all of the suppliers came at it independently, with no standardized set of technologies. For 100Gbit/s, the industry took a much more planned approach with which operators are now far happier (based on Heavy Reading surveys). This process must be repeated for the next-generation transport rate, and this planning takes time.
With network traffic continuing to increase, a transport rate beyond 100Gbit/s will be needed someday. Work needs to start sometime, and, with the major work on 100Gbit/s completed, this seems like a reasonable time for the work to begin.
— Sterling Perrin, Senior Analyst, Heavy Reading