Technical sessions at the 35th annual European Conference on Optical Communication (ECOC), one of two key annual optical communications conferences, were heady with both long-term visions and recent experimental results.
The focus was on the latest exotic, high-speed modulation schemes and switch technologies for terrestrial and on submarine transport networks and different approaches for “next-next-gen” access networks. But this analyst walked away with an overarching concern: will the telecoms industry be able to get out of its own way to generate a return on investment in 40G while also paving the way for profitable 100G products to come?
Taking advantage of 40G/100G opportunity, as distilled by the week of formal presentations and the informal talks with component and systems vendors and network operators, boiled down to three main industry approaches:
- The hares approach suggests that 100G, for both short-reach (e.g., data center) and long-reach (e.g., metro and backbone) applications, is technically mature and will provide an economical alternative to both nascent 40G and today’s volume leader, 10G, by 2011 or so. The hares tend to be either vendors that are late to the 40G party or operators that make “if I had it I’d install it today” comments to gullible vendors, which in turn produces more hares.
- The tortoises include those that have or think they have figured out how to make money on 40G - and this author. We see 100G as realistically reaching commercial costs and volumes beyond 2012, given remaining technical challenges and the fragility of the optical food chain. This group views 40G and 100G as conjoined twins that could both die if separated. First, it’s essential for the health of the component and system vendors to reap revenues from the significant 40G R&D they’ve sowed over the last decade. Second, much of the technology can be shared between 40G and 100G.
- The deer are so worried about making a wrong move and squandering millions in development dollars on a phantasm that they’re effectively inert.