Trends and evolution of optical networks and technologies

Research: exploit technology cross-fertilization

Optics

The fundamental trends in telecommunications
- more bandwidth hungry services, more
intelligent and easy to manage networks - will

M. Erman

inevitably bring optical technologies back
into the spotlight.

Trends and evolution of optical networks
and technologies

Introduction

and WDM. During 2000, the volume of data traffic in the
USA surpassed the volume of voice traffic, highlighting the
dominant role that the Internet is playing in terms of services. Forecasts indicate that growth in data traffic will continue at an exponential rate in the years to come.

Over the past few years, optics has established itself as
one of the basic communication network technologies as
a result of the conjunction of several key technological
innovations (optical fiber, semiconductor lasers, fiber
amplifiers) and market needs. Thanks to the introduction
of Wavelength Division Multiplexing (WDM), optical
transmission now makes it possible to transmit enormous
amounts of information over almost unlimited distances.
As far as transmission capacity is concerned, fiber has no
competition. In addition, optics offers a number of
advantages in the field of networking. Even though
recent cuts in capital expenditure (capex) have slowed
down progress in this field, the fundamental trends in
telecommunications will inevitably bring optical technologies and networks back into the spotlight.
Optical communications is still a very new industry. Fibers
have only been widely installed over the past decade, and
mainly for long distance transmission. Optical networking
is not really here yet. The industry is young, and consequently somewhat immature. Hence, research in optical
communication technology can actively contribute to
improving the technology in various industrial as well as fundamental areas, such as materials, devices, architectures
and protocols.
This article examines the general trends in optical communications and describes Alcatel’s main research directions. Some of the key Alcatel research results are highlighted in other articles in this “ Optics ” section of the
Alcatel Telecommunications Review.

One of the primary objectives of optical research is therefore to pave the
way for increased capacity and more
intelligence in optical networks.
Meanwhile, as operators were investing to cope with Internet growth (although revenue was still coming primarily
from voice traffic) the capital crunch occurred. This
increases the importance of another research objective:
make the technology less expensive to produce, implement and operate.
To prepare for the future, innovation in the field of high
capacity transmission remains important. Long haul systems – both terrestrial and submarine – go through the well
known “build and fill” cycles. Industry-wise, we are in a “fill”
cycle. From a research perspective, it is important to prepare for the next “build” cycle. The next generation of optical transmission will be based on N x 40 Gbit/s systems, providing capacities of several Tbit/s. However, when considering transmission, one has to take into account the distance between regenerators because the Signal to Noise
Ratio (SNR) degrades with distance. Consequently, transmission needs to be evaluated with respect to both capacity and distance: For example, future transmission systems
might be identified as “10 Petabit/s*km networks”, meaning that they offer 10 Tbit/s transmission over 1000 km, or
1 Tbit/s over transoceanic distances.
Alcatel is strongly committed to N x 40 Gbit/s systems.
Research embraces all the relevant fields, including
submarine and terrestrial systems, as well as a range of
enabling technologies. For example, Alcatel was one of
the first to demonstrate transmission at more than
10 Tbit/s [1]. A number of innovations made this demonstration possible: vestigial sideband filtering at the
receiver for narrow channel spacing, distributed Raman
amplification to optimize the SNR, Alcatel Teralight
fiber to reduce fiber impairments, and polarization multiplexing to double the capacity.

More Bits to More Users
The explosive growth in capacity is largely a result of massive use of the Internet. The combination of an increasing
number of Internet users and the introduction of new content-richer services with more picture and video content
has resulted in the demand for capacity doubling every 6
to 9 months in some networks. Such growth – faster than
was experienced in electronics – has been possible thanks
to the combination of Time Division Multiplexing (TDM)

Alcatel Telecommunications Review - 3 rd Quarter 2001

173

Trends and evolution of optical networks and technologies

As a result of these innovative technologies, Alcatel
achieved a record spectrum density of 1.28 bit/s/Hz. This
parameter is important as it indicates the efficiency of
spectrum utilization and is therefore linked to the cost.
The achieved efficiency is six times higher than for today’s
commercial systems. In another experiment, N x 40 Gbit/s
transmission was demonstrated in a submarine configuration. Transmission at 32 x 40 Gbit/s (in excess of 1 Terabit/s) was achieved over a distance of 2400 km using
amplification only, and no regeneration.
In the case of ultra-long-haul transoceanic systems,
N x 40 Gbit/s systems will require regeneration. Although
one can implement this function using optoelectronic conversion, this would be a step backwards compared with
the present situation in which one optical amplifier is used
to simultaneously amplify several (in most cases all) wavelength channels. An optoelectronic regenerator is, by definition, a single-channel device that might jeopardize the
cost advantage optics has brought to transmission. Thus
research into optical regenerators is a key program. Alcatel is investigating several approaches based on semiconductor wavelength converters, in-line synchronous
modulation and saturable absorbers [2].
Optical transmission on long haul networks is only part
of the picture. Fiber will inevitably be the transmission
medium in metropolitan area networks, and is increasingly
extending its reach into access networks. Following the
generalization of high speed Internet accesses (Asymmetric Digital Subscriber Line, ADSL; Very high speed
Digital Subscriber Line, VDSL; etc), a need will soon
emerge for high capacity transmission systems to the customer premises. As a result, photons are coming closer
to the home! However, metropolitan and access networks
raise a number of challenges other than purely transmission ones: protocols, multiservice capability and cost
are the dominant issues.

detecting only a small percentage of the signal. In collaboration with European partners, Alcatel has demonstrated an all-optical cross-connect and has tested it in
a real network [3].
Optical switching can, however, find a place even in a
network that is not fully transparent. Indeed, an electronic switching matrix can be replaced by an optical
one. In this case it does not provide a specific functional
advantage, but the expectation is that for large switching matrices, an optical implementation will be cheaper
than an electronic one. An optical switching element is
also bitrate independent, which means that it is possible to upgrade ports from, say, 2.5 Gbit/s per channel
to 10 Gbit/s, or even 40 Gbit/s, without changing the
matrix. This is not possible with an electronic version
since higher capacity requires more processing power.
Whatever implementation is selected, such cross-connects perform wavelength switching, and thus allow
wavelength service (end-to-end wavelength provisioning, for instance). Signaling, controlling and managing WDM networks have become hot research and
development topics.
Besides the introduction of an Optical Channel (OCh),
which makes it possible to treat each wavelength as a
separate logical channel, Internet-based protocols,
such as Multi Protocol Wavelength Switching (MPλS)
and Generalized Multi Protocol Label Switching
(GMPLS), are being introduced at the control layer. The
basic driving forces are known: apply data-oriented protocols (which proved so cost-effective for the Internet)
to WDM networks and make dynamic establishment of
wavelength-based routing paths possible. As data is
becoming the dominant type of traffic, this trend
appears natural. Nevertheless, the required constraints
on Quality of Service (QoS), restoration and protection
need to be carefully assessed.
The impact of data is even larger and more profound on
metropolitan networks. Because of the mix of different
formats – Internet Protocol (IP), Asynchronous Transfer Mode (ATM), Gigabit Ethernet, etc – such networks
naturally have to evolve towards multiservice networks.
On the optical layer level, WDM is the most suitable technology, yet with even greater cost constraints. Transparency, which is difficult to manage at the backbone
layer, might find an easier implementation in the
metropolitan area. Alcatel research is working on a number of innovative solutions [4].
The ultimate dream of an “IP-over-optics” approach
remains, however, an optical router. This requires fast
optical switching fabrics. Alcatel already has considerable experience in optical packet switching, having
demonstrated the first optical ATM switching demonstrator some years ago as part of the European ATMOS
and KEOPS programs. We have further refined these
ideas and have adapted the concept to take into
account the IP dimension. The first burst optical
router has been assembled; it exploits a number of innovative approaches for both the optical elements and the
control layer. This prototype has validated the feasibility
of implementing an all-optical burst router – including
burst transmitters and receivers – as well as high-speed
scheduling algorithms.
It is clear that optics can offer much more than just pointto-point transmission. Wavelength service, network pro-

From Dumb Pipes to Intelligent Networks
If optical transmission – and WDM in particular – has
established an undisputed leadership, the use of photonics and exploitation of the wavelength domain for networking is still in its infancy. Nevertheless, it is tempting to push further what photons can do in a network.
The argument is simple. Consider a WDM network with
80 wavelength channels of 10 Gbit/s on each fiber. On
each of the network nodes, a cross-connect will have to
switch hundreds of 10 Gbit/s channels from input fibers
to either drop channels or output fibers. Electronics is
the way to do it today. However, this requires a
transceiver – which involves optoelectronic conversion
– at both the input and output ports. These transceivers
are the major cost element in a cross-connect.
As most of the traffic in a node is transit traffic, replacing the electronic cross-connect by a fully transparent
optical cross-connect is the obvious “low cost” photonic
alternative. It does, however, raise a number of issues,
including the non-intrusive monitoring required to manage all-optical networks. Several solutions are being
investigated within the Alcatel laboratories. These solutions are either based on additional control channels or
modulation, or the use of high speed electronic processing capable of assessing the quality of the signal while


174


tection at the optical layer, wavelength routing and, eventually, a “true” IP-over-optics implementation are some
of the evolutionary steps that are at an advanced stage
within Alcatel Research.

wider temperature range (from –40 to +85°C) - an
attractive challenge for quantum mechanics specialists!
The next move was to replace active fiber/laser alignment
by a passive technique. This was achieved by rethinking
the laser mounting process. Silicon motherboards have
been developed, which make it possible to use an automatic self-aligning process for the laser and fiber, with the
help of “indentation” and appropriate structuring of the
laser chip. However, further innovations were needed at
the laser chip, such as the integration of a taper (the equivalent of an integrated lens).
All this was necessary in order to develop the surfacemountable plastic laser modules. Nothing would have been
possible without innovation in various fields of physics,
optics and processes.
What is coming next? One trend will clearly be to integrate
more functions, including both passive optical functions
and dedicated electronic interfaces. SiO2/Si motherboards will play a key role in assembling the passive and
active optical parts cost-effectively. There are a number
of other interesting options for WDM components [6].
These trends indicate that the components

Components: a Pace of Change
Components need to meet two sets of objectives: one concerns their performance and function, while the second
is linked to cost constraints. Both explain why optical components are at the heart of today’s communication systems. Not only do they set the performance limits and
functional constraints, but also, as they represent a significant and increasing proportion of the equipment
cost, they strongly impact the final system cost.
The trends mentioned above for high speed transmission
and intelligent networks will materialize only if suitable
technologies are available. The research highlights
included in this issue of the Alcatel Telecommunications
Review show where the challenges for optical components lie from a functional point of view. Transmission
at 40 Gbit/s requires high-speed modulation, detection
and associated electronics. Managing fiber impairments
(chromatic dispersion, polarization mode dispersion,
etc) requires dedicated passive components. Dense
WDM requires multiplexers and demultiplexers for
higher channel counts and narrower channel spacings.
Dedicated electronics, particularly the stages that interface directly with the optoelectronic chips, will be
equally important for high-speed systems. Optical amplification needs to be developed for new wavelength windows (after C and L, the next window will be S), while
at the same time the increasing number of channels will
require more power [5]. Other functions become mandatory when moving towards optical networks: optical
switches of course, but also devices capable of monitoring
the QoS, and ultimately, optical regenerators. Alcatel
research has achieved breakthroughs in all of these fields.
As regards cost, one might think that this is more an industrial than a research issue. In fact, the cost of optoelectronic components has been reduced, and will continue
to be reduced, through innovation.
Oversimplifying, we can say that an optoelectronic device
is made of a chip (front-end) packaged in a module (backend). The short history of evolution of optoelectronic
devices was an alternation of breakthroughs in the frontend and back-end processes.
The first important step was at the beginning of the 90s
when, for the first time, Alcatel demonstrated the feasibility of manufacturing full 2 inch InP wafers, each with
15 000 lasers! This was made possible thanks to the development of strained quantum well lasers in the research
laboratory. The technology proved capable of producing
high-performance lasers which were uniform and reproducible. It also represented a breakthrough in the cost of
the laser chip.
However, the dominant cost then became the module,
which was metallic, used a Peltier cooler and needed very
accurate (manual) fiber/laser chip alignment. The first step
was to eliminate the Peltier cooler and develop the socalled coax module, still using active alignment. Again this
made it necessary to go back to the laser chip and
develop new laser structures that could operate over a


industry will evolve considerably over
the next few years to offer ever higher
performance and, even more important, greater functional integration at
lower costs. Thus the industry will progressively
mature. Research and innovation are important factors
in making this happen.

Innovation: the Art of Networking
So far, we have focussed on the near- and medium-term
evolution of optical technologies. However, optics is

here to stay for a long time, and disruptive technologies will inevitably
appear. Some may already be knocking at the door,
for example, photonic bandgap materials which might
be used for optical fibers, planar passive devices and
semiconductors. In a world where innovation can happen in various places and environments, where a real
application might be difficult to detect at an early stage,
in other words, in a world of uncertainty, how should we
manage innovation? Alcatel believes that partnership is
the right way to go; it can take various forms.
Consider some examples in the optical field. Multi-partner projects – national and international – make it possible to build efficient multidisciplinary projects combining the talents, expertise and vision from universities and industry. In Europe, Alcatel is a major player
both within national projects, such as BMBF in Germany and RNRT in France, as well as within international projects, such as IST. Many of our advanced studies in the area of optical networking have been initiated in this context, and projects such as OPEN
(transparent optical cross-connect), KEOPS (optical
packet switching), MEPHISTO (management of alloptical networks) and PELICAN (field trial implementation of an all-optical network) were the first to
explore new, innovative options.

175


Fig. 1 Some of the key participants in “Optics Valley”
In the field of basic technologies, Alcatel is also very committed to partnerThales Central Research
ships through international projects.
Ecole Polytechnique
Projects on advanced topics such as
University Paris-Sud
Orsay
CNRS - “LULI, LOA”
IEF
photonic bandgap materials, quantum
Supelec
Z.I. Courtaboeuf :
boxes and, more generally, nano-techIOTA
Picogiga
…
...
nologies, are areas of ongoing activity.
In specific fields, bilateral cooperation
Alcatel Optics
can further help to achieve impressive
Terrestrial & submarine
progress for the benefit of both parties.
transmission
OPTO +
Alcatel R&I
With this in mind, Alcatel has launched
and is supporting a number of collabCNRS-LPN
orations with major Universities and
Institutions worldwide. As an example,
some remarkable results have been
achieved in a bilateral program with
the Heinrich Hertz Institute, one of our key partners [7].
research teams. Some other elements will require disAnother such initiative is the creation of “Optics Valley”.
ruptive approaches that may not yet have been identified.
Located south of Paris, Optics Valley is an association of
This is where cooperation with more academic centers of
major universities, engineering schools, small to medium
excellence will play a determining role.
size enterprises, and large corporations that are active in
Open your eyes and let the light come in! s
the optics field (see Figure 1). It represents a unique pool
of skills in both fundamental and applied sciences. Alcatel was one of the founders of Optics Valley and is the refReferences
erence industrial partner working on optical communication. The various participants in Optics Valley are
1. S. Bigo, W. Idler, A. Scavennec, L. Du Mouza: “Road to
expected to give birth to many promising university/indusultra-high-capacity transmission”, Alcatel Telecommunicatry collaborations. As an example, a prestigious CNRS labtions Review, 3rd Quarter 2001 (this issue), pp 177-178.
oratory working on optics and nano-technologies (LPN)
2. F. Brillouet, F. Devaux, M. Renaud: “From Transmission
will be collocated with the Alcatel research laboratory in
to Processing: Challenges for New Optoelectronic
Marcoussis. The exchange of ideas and the collaboration
Devices”, Alcatel Telecommunications Review, 3rd Quarter
facilitated by the proximity of two large laboratories – one
1998, pp 232–239.
with an industrial culture and missions, the other with
3. J. L.Beylat, M. W. Chbat, A. Jourdan, P. A. Perrier: “Field
more fundamental objectives – will certainly foster innoTrials of All-Optical Networking based on Wavelength
vation. Thus, we believe that cooperation,
Conversion”, Alcatel Telecommunications Review, 3rd
partnership and networking with other
Quarter 1998, pp 218-224.
centers of excellence are important
4. A. Jourdan, L. Tancevski, T. Pfeiffer: “How much optics
in future metropolitan networks?”, Alcatel Telecommunicaassets. After all, it is interconnection that provides
tions Review, 3rd Quarter 2001 (this issue), pp 219-221.
intelligence to the human brain!
5. D. Bayart, L. Gasca, G. Gelly: “Cladding-pumped erbiumdoped fiber amplifiers for WDM applications”, Alcatel
Telecommunications Review, 3rd Quarter 2001 (this
Conclusion
issue), pp 179-180.
6. J. Jacquet, K. Satzke, I. Riant: “Low cost DWDM
We have reviewed some of the trends in optical networks
devices”, Alcatel Telecommunications Review, 3rd Quarter
and technologies. Although optical telecommunications
2001 (this issue), pp 181-182.
now appears to be a well-established technology, it has
7. F. Devaux, O. Leclerc, B. Lavigne, P. Brindel, H. P. Noltreally only been extensively used for transmission for
ing, B. Sartorius: “Alcatel-HHI collaboration on all-optiabout ten years. Many challenges and opportunities are
cal 3R regeneration”, Alcatel Telecommunications Review,
ahead of us. The future is only partly predictable.
3rd Quarter 2001 (this issue), pp 231-233.
Increased capacities are inevitable, even if the present
economic slowdown might change some of the milestones. The move to intelligent optical networks is also
a strong move which will give added value to operators.
Marko Erman is Senior Research & Innovation
Many of the advances needed to implement this vision are
Director and Member of the Optics Group Board.
already in the laboratory, as is illustrated by several artiHe is based in Marcoussis, France.
cles in this issue of the Alcatel Telecommunications
Review. These articles also demonstrate the strong
commitment and quality of the results of Alcatel’s


176


Trends and evolution of optical networks and technologies

More Related Content

What's hot (20)

Similar to Trends and evolution of optical networks and technologies (20)

More from Md.Bellal Hossain (6)

Recently uploaded (20)

Trends and evolution of optical networks and technologies