William Shieh talks with
ScienceWatch.com and answers a few questions about
this month's Fast Moving Front in the field of
Physics.
Article: Coherent optical orthogonal frequency
division multiplexing
Authors: Shieh,
W;Athaudage, C
Journal: ELECTRON LETT, 42 (10): 587-589 MAY 11 2006
Addresses: Univ Melbourne, ARC, Special Res Ctr
Ultrabroadband Informat Networks, Dept Elect & Elect
Engn, Melbourne, Vic 3010, Australia.
Univ Melbourne, ARC, Special Res Ctr Ultrabroadband
Informat Networks, Dept Elect & Elect Engn,
Melbourne, Vic 3010, Australia.
Why do you think your paper is highly
cited?
For the first time in "open access" literature, this paper introduced a
novel modulation format which combines two powerful techniques in optical
communications, coherent detection, and orthogonal frequency-division
multiplexing (OFDM).
This new modulation format called coherent optical OFDM (CO-OFDM) holds the
promises of delivering high electrical and optical spectral efficiency,
receiver sensitivity, and optical dispersion resilience. As such, CO-OFDM
has emerged as one of the attractive candidates for the forthcoming 100
Gb/s and 1 Tb/s Ethernet transport.
Does it describe a new discovery, methodology, or
synthesis of knowledge?
It describes a new approach of combating optical chromatic dispersion (CD)
and polarization mode dispersion (PMD) in optical communications.
Independently from other research groups in Monash University, Australia,
and the University of Arizona in the USA, who looked at the problem from an
optical direct-detection side, we applied coherent optical OFDM to
long-haul optical communications, which presents an exciting potential for
future ultrahigh-speed optical networks.
Would you summarize the significance of your paper
in layman's terms?
Internet traffic has been growing at an alarming rate of 40-50% per annum.
This places significant strain on the underlying communication
infrastructure, which is predominately based on fiber optical transport.
The line rate of the optical signal has increased from 2.5 Gb/s to 40 Gb/s
in the last decade, and now 100 Gb/s is the speed being asked for by
carriers and developed by suppliers.
"Our society has entered an
information age that is heavily centered
around the internet. As such, it is
imperative that the underlying infrastructure
is able to meet the demands from the
relentless internet growth stemming from
bandwidth-rich applications such as YouTube
and Internet Protocol Television
(IPTV)."
However, the ultrafast optical signals are very sensitive to the optical
dispersion in the fiber, and subsequently managing the dispersion to allow
for 100 Gb/s has become a difficult problem. The analogy can be made
between this phenomena and a Formula 1 car race, a small bump on the track
can easily wreak havoc on a highly specified and fast moving race car.
OFDM provides a smart solution to this problem. In layman's terms, it is a
"divide-and-conquer" strategy. OFDM partitions the entire channel into many
small segments. Even though the optical channel is relatively uneven in its
entirety, each segment is quite smooth and flat, and subsequently can be
managed relatively easily. It is the same reason why OFDM has been very
successful in wireless communications.
There is one more piece of the puzzle that needs to be uncovered in order
to apply OFDM in optical systems. OFDM consists of many orthogonal
subcarriers, each essentially an oscillating wave with its amplitude
swinging up and down between positive and negative values. One of the
stumbling blocks to apply OFDM is that conventional optical communication
systems are based on intensity modulation which, by nature, is positive.
This makes application of OFDM unnatural, with inherent sacrifices to
either optical power efficiency or spectral efficiency.
This is where the advantages of coherent detection come into play; with
coherent detection, both positive and negative signal amplitudes are
permitted. However, coherent detection has more components and thus is more
costly.
Fortunately, coherent detection enables dual polarization transmission,
doubling the capacity and thus alleviating the cost-disadvantage.
Furthermore, dual polarization transmission completely overcomes the
polarization mode dispersion impairment, which has been considered as the
fundamental barrier for ultrafast optical communications.
The introduction of CO-OFDM could prove opportune, as the required digital
signal processing power can be provided either today, or in the near
future, by the rapidly advancing powerful silicon technology underpinned by
the well-known Moore's Law.
How did you become involved in this research, and
were there any problems along the way?
I have performed extensive research in optical communications for 16 years.
I have been doing research with my Ph.D. students on coherent optical
communications for the last five years. At the time, the
finite-impulse-response (FIR) filter and the Viterbi-type
Maximum-Likelihood Sequence Estimator (MLSE) were two approaches we were
studying to mitigate the optical dispersion.
I was frustrated with the computation complexity of both FIR filters and
Viterbi decoders. In a casual conversation with one of my colleagues in our
department, Dr. Chandra Athaudage (the other co-author of the paper), he
mentioned to me that he was working on OFDM, which was a popular modulation
technique to combat multipath interference in wireless systems.
I began to study the OFDM literature, and felt that I had found the
solution for the optical problems at hand. I laid out the necessary
equations and ran several simulations, and found the results were quite
impressive. These were later submitted to Electronic Letters.
In the field of optical communications, experimental demonstration is
always regarded as a premium. We embarked on the first proof-of-concept
experimental demonstration for CO-OFDM. It turned out to be one of the most
difficult experiments I have ever attempted.
CO-OFDM is quite a complicated system involving non-trivial experimental
configurations and signal processing. At the beginning, we tried very hard
and could not recover the CO-OFDM signal properly. My students and I were
constantly pondering about whether CO-OFDM simply did not work, or whether
we simply had some bugs in the hardware or software. There were many times
we thought we ran into a dead end and came very close to the conclusion
that CO-OFDM was just a theoretical fantasy.
Finally, we triumphed and managed to demonstrate a multi-Gigabit CO-OFDM
transmission. Luckily we were not alone in the demonstration game. After
our first CO-OFDM demonstration, progress has been accelerated by
tremendous efforts from other institutions such as Monash University in
Melbourne, Australia (with direct-detection optical OFDM); Japan's KDDI
R&D Laboratories and Nippon Telegraph and Telephone Corporation's (NTT)
Photonics Laboratories, along with Alcatel-Lucent's Bell Labs in Stuttgart,
Germany and New Jersey, USA, who are all in a friendly race to outstrip
each other with better results. Many "heroic" experiments in optical
communications are produced by using CO-OFDM.
Do you foresee any social or political implications
for your research?
Our society has entered an information age that is heavily centered around
the internet. As such, it is imperative that the underlying infrastructure
is able to meet the demands from the relentless internet growth stemming
from bandwidth-rich applications such as YouTube and Internet Protocol
Television (IPTV).
There is a great concern for the sustainability of such explosive growth in
a cost-effective manner. Emerging new technologies such as CO-OFDM provide
an important step in the right direction for the next generation of optical
networks by facilitating higher spectral efficiency and dynamic bandwidth
provisioning, in order to reduce the overall installation and maintenance
costs.
William Shieh, Ph.D.
Associate Professor and Reader
Department of Electrical and Electronic Engineering
The University of Melbourne
Victoria, Australia