Smitten with Nature’s Bounty
Below you’ll find an excerpt from the June issue of The George Gilder Report. If you’re not already a subscriber, click here to learn more information on how to gain access to this research and the recommendation.
There are several reasons why I was on the side of optical.
Sending photons gliding across continents almost effortlessly on pathways of the purest glass, free from time-consuming, costly power-hogging electronic intervention, was efficient. When the first optical networks appeared, they seemed an almost fantastical instance of nature’s bounty.
From the outset, single optical fibers could transport more than a thousand times as much information per second as the most capacious copper phone lines. Within a few years, single fibers, bearing up to 160 wavelengths, were carrying terabit loads (trillions of bits per second), making them a million times faster than their copper counterparts.
Supremely elegant, the all-optical network was organic, non-GMO, and all-natural. It was fully defined by the physics of light, a gift of nature and nature’s God. Even the switches and multiplexers were all natural: sophisticated versions of devices that might be found in any middle-school science room — mirrors and prisms, admittedly of fantastically small scale.
Who could resist its charms?
And once smitten, how could I fail to defend my lady love from the deformities imposed on her by all those thermodynamic electronic machines? Yet, at the very time I was defending the purity of photonic bandwidth. I was celebrating Qualcomm for its man-made, computational, Moore’s Law-driven solutions to nature’s shortcomings. Even then, I knew that the notion of bandwidth as a pure gift of nature is always an illusion.
There is no bandwidth anywhere until man makes it by making waves. Ah, well. I was blinded by the light.
Photons, meet Wall
The illusion of the all-optical network collapsed about a decade ago, when the photons hit a wall. Any signal, from a tom-tom to a modulated wavelength of light, is subject to attenuation and distortion over time and distance. In particular, the more signals per second, the harder it is for the signal to get through clearly.
Think about it… If a drummer beats the drum too many times per second, you will be unable to distinguish one beat from another. The effect becomes worse with distance, as sound waves echo and ricochet so that the vibrations from the same beat of the drum hit your ear at different times.
The optical version of this effect is called chromatic dispersion, one of several distortions to which optical signals are subject. Conventionally, when describing an optical transmission, we say we are transmitting over a given wavelength or frequency of light, say 1530 nanometers (nm). In reality, that single number stands for a band of wavelengths working together to give the signal sufficient power to reach its destination. In a vacuum, all would travel at the speed of light.
In any medium, however, including even the wonderfully pure glass of fiber optics, the various wavelengths in our band will travel at minutely different speeds.
Over time, the fastest wavelengths in the band catch up to the slower wavelengths bearing the bit just ahead of them. As a result, the signal is distorted. Several factors contribute to chromatic dispersion. One is distance. Another is the frequency of the transmission.
Faster signals, more pulses per second, require a “wider” swathe of spectrum. Roughly speaking, transmitting at 10 billion pulses per second (10 gigabaud) will require a bit more than 10 gigahertz (Ghz) of spectrum.
The wider the swathe of spectrum and the greater the difference in speed between the fastest and slowest photons, the more likely it is for the fastest photons carrying one bit to pass the slowest photons carrying another. Now, for an all-optical network, chromatic dispersion was able to be overcome by purely physical adjustments to the composition of the fiber itself.
But not anymore. Physical adjustments were effective only up to about 10 gigabaud, depending on the distance the signal had to travel. We hit this limit just over a decade ago. Nature, it appeared, had given all it had to give.
To get beyond 10 gigabaud (which in those days translated directly to 10 gigabits, with each pulse of light carrying one bit), nature would need computers. And computers need electrons.
The reign of the optoelectronic network had begun.
To read the rest of the research and learn more about the monthly recommendation in The George Gilder Report — if you’re not already a subscriber — click here.
Editor, Gilder’s Daily Prophecy