The charm of history and its enigmatic lesson
No CommentsPosted by simonpoole on December 16, 2009
As this marks my first ‘official’ post for my new column, I’d like to start by looking back over the past 30 years that I’ve been involved in the optical communications industry to see whether Aldous Huxley’s aphorism that “The charm of history and its enigmatic lesson consist in the fact that, from age to age, nothing changes and yet everything is completely different” holds true for our industry.
Whilst it’s hard to think of something as recent as optical fiber communications moving from “…age to age…”, it can, however, be viewed as a series of technological waves – each of which builds on the momentum and advances of the previous. Once Kao and Hockham’s work had identified silica as the optimum transmission medium, and the group at Corning proved that low-loss fibers could indeed be made from silica, the challenge was then to develop the communications systems based on this wonderful new material.
With the benefit of 20/20 hindsight, each of the past three decades in optical communications can be characterized by a significant technological change and, as we shall see, I think we are about to embark upon yet another revolution in the next few years.
The 80’s
The time of The Smiths but also big hair and Maggie Thatcher – and telecoms made the transition from multimode to single-mode fibers. Single-mode fiber now dominates the telecoms transmission part of our industry but multi-mode fiber has not gone away and is still the dominant medium for client-side optical transmission.
The 90’s
In the 1990s the shift was from single wavelength SDH/SONET at 1310nm to DWDM at 1550nm using EDFAs. Whilst the first commercial deployment of EDFAs in 1993 occurred the canonical seven years after the invention of the EDFA in 1986, it looks like it will only be next year (2010) that world-wide shipments of amplified DWDM systems finally overtake those of single-wavelength (SONET/SDH).
Present Day
As of late 2009, we are in the middle of the transition from fixed network architectures to reconfigurable architectures. This has come through a progression from Wavelength Blocker technology through PLC-based ROADMs to today’s weapon of choice for reconfiguration – Wavelength Selective Switches (WSS). Again, optical reconfiguration has been around for nearly a decade now, but reconfigurable systems, whilst being the most rapidly growing part of the market, still account for less than half of all WDM system sold.
The Future
The next wave to hit and one which, I think, has fundamental implications for our industry, may be that of coherent transmission. Proposals for coherent transmission systems have been around since the early 1960s. Coherent transmission even had a brief flowering in the early 1990s before being steamrollered by the oncoming EDFA/DWDM juggernaut and put into the “interesting research but doesn’t really work for optical communications so we’ll call it a sensor” basket. It was then forgotten for a decade or more whilst we all got busy riding the photonics bubble of the late 1990s. In the past few years, however, the seemingly irresistible advance of Moore’s Law has raised the capabilities of silicon processors to the point where we can now dust off once again the potential benefits of coherent communications, add to them the processing power of an ASIC and finally drag optical communications kicking and screaming away from spark-gap transmitters and into the radio era circa 1930 (albeit with a slightly higher bandwidth).
Once again this change could radically alter the operation of our industry, with a move away from purely optical capabilities and an increasing reliance on silicon (as Physicist and Nobel Laureate Arnold Penzias once said, “If you find yourself fighting Si, don’t” ). However, as in previous changes, the adoption of coherent transmission likely will initially take place at the bleeding edge of the system, at 40 and 100 Gb/sec. However, given the additional cost and optical complexity of a coherent transceiver (i.e. additional modulators, local oscillator, etc.) coherent transmission is unlikely to displace existing technologies at 10G and below. More likely coherent will be the “weapon of choice” for future evolution to even higher per-channel bitrates (1 Tbit/sec/channel, anyone?) which provide improved immunity to the non-linear Shannon limit.
So, does Huxley’s postulate hold true? Does nothing change and yet everything is different?
From a technical perspective, everything is indeed different. Each wave of technology has brought huge changes to our industry and, as a result, immense increases in the ability of people to access information when and where they need it.
On the other hand, each wave follows roughly the same timeline for adoption, taking at least a decade to displace the then incumbent technology (I nearly used the words “paradigm shift”, but that’s for a future post) and in that regard, nothing changes.
Finally, the last word this week should go to JCR Licklider, one of the grandfathers of the Internet and founder (in 1962) of the Intergalactic Computer Network, who, after many years of watching how technologies are developed, hyped, and then finally adopted noted that: “People tend to overestimate what can be done in one year and underestimate what can be done in 5 or 10 years“.
