Multiplexing and the Capacity Crunch

Chris Goldenstein
June 10, 2012

Submitted as coursework for PH250, Stanford University, Spring 2012


Over the last 25 years, fiber optic networks (FON) have become an increasingly popular and powerful advancement in the telecommunications industry. Fiber optic networks offer several advantages over traditional cable networks. Perhaps the most important advantage of using FON is the increased bandwidth and data capacity. FON also enable the transmission of cleaner signals as they are not susceptible to many of the noise sources that can interfere with electrical signals (e.g. electro-magnetic interference). However, as our appetite for bandwidth hungry activities continues to grow, even FON are struggling to meet our bandwidth needs. At the heart of this problem lies the need to transmit more and more information over a given carrier medium (e.g. cable or optical fiber). Wavelength-division multiplexing (WDM) has been one of the most popular methods for optical signals, however, this technique is limited. Several alternatives to WDM exist, however, one attractive variant of this technique is spatial multiplexing or space-division multiplexing (SDM) which may enable the delay, but not prevention, of the so called "capacity-crunch."

Fundamentals of WDM

In WDM, multiple colors of light are combined onto a single optical fiber. Each wavelength of light carries its own signal and as a result, signals can be sent across both directions of the fiber. In the simplest case, the signals are demultiplexed on the receiving end using spectral filters that separate the various colors of light into separate paths for detection. In dense wavelength-division multiplexing (DWDM) there can be dozens of signals multiplexed onto a single optical fiber with up to 18 infrared-channels and each IR channel carrying its own set of frequency-division multiplexed (FDM) signals. The individual IR channels are separated via spectral filtering while their respective FDM signals are separated using Fourier-transform based signal processing techniques. Currently, commercial WDM based systems can transmit 10 Tb/s with approximately 100 Gb/s per wavelength. Research grade systems have reached the 100 Tb/s per fiber mark, however, significant increase beyond this mark is not expected. [1] Growth in WDM system capacity has slowed from 80% per year in the 1990s to 20% per year since 2002 and at the current demand growth rate, systems capable of exceeding the Shannon limit will be needed in the near future. This realization has led to the dawn of the so-called "capacity crunch" and a need for new multiplexing techniques [1].

Potential of SDM

One attractive option to overcoming the capacity crunch is to implement SDM into fiber optic networks. In its simplest form this simply consists of employing parallel optical paths from source to destination. While this technique is currently not economically feasible since it does not reduce the cost per bit, it is expected to become economically attractive as further improvements in spectral efficiency become unsatisfactory or fundamentally limited. For example, using today's most spectrally efficient systems, a spectral efficiency of 20 b/s/Hz would require nearly three orders of magnitude more transponders than if a more standard multiplexing approach was taken with three parallel optical paths. [1] With that said, SDM clearly offers potential for affordable high bandwidth networks capable of fulfilling our growing bandwidth needs.


Optical network traffic has grown at 30-60% per year since 1991. [1] This exponential growth in demand has led to the need for advanced information multiplexing techniques. One common strategy is WDM and variants thereof, however, this strategy is fundamentally limited and will be unable to fulfill our growing needs for higher data transmission rates. One strategy that offers potential for higher bandwidth optical networks is SDM where multiple optical paths can be laid between destination and source, thus scaling the system capacity linearly. While this technique is costly, as every optical path scales the system cost, it may become more cost effective as more expensive and advanced multiplexing techniques are needed to meet the growing demand for higher data transmission rates.

© Chris Goldenstein. The author grants permission to copy, distribute and display this work in unaltered form, with attribution to the author, for noncommercial purposes only. All other rights, including commercial rights, are reserved to the author.


[1] P. J. Winzer, "Optical Networking Beyond WDM," IEEE Photonics J. 4, 647 (2012).