NTT R&D plans (1994). part 2 of 2

[David Farber <farber () central cis upenn edu>]

new services.  "Access" may involve the use of optical fibers or radio
means -- in particular, a variety of approaches to the former method are
currently being developed on an international scale.  At NTT, we are
investigating the construction of a highly reliable and economical
optical access network based on the PDS (Passive Double Star)
architecture.  This section introduces some of our experimental work in
this area.


The fiber-optic access network at which we are aiming will be introduced
into offices and all kinds of areas, and then into the home, enabling
customers to use a variety of services through one strand of optical
fiber.  It should also be possible for services to be added and updated
without delay.  To meet these requirements, the access network must be
capable of establishing any number of circuits between the customer and
multiple service nodes as needed.  With this in mind, NTT proposes an
"access platform" that offers wide support for all services and can be
applied to B-ISDN as well.


The access platform is basically a mechanism for establishing links with
the required number of circuits of the required capacity.  It uses PDS
transmission and multiplexing techniques to freely manipulate "paths"
(the routes taken by circuits) through an optical link network.


The platform is currently based around STM technology, but we expect
this will be replaced by ATM multiplexing techniques which have superior
flexibility.  Work carried out to set up the required number of circuits
of the required capacity for each customer will be carried out entirely
as remote operations by the operations center.  As customers come to
take full advantage of the fiber-optic access network based on these
concepts, we can expect services to be used in more effective ways and
more powerful software to be developed.




5.  Intelligent Layer


The potential value of a network rises sharply as the quality of the
services it provides is increased.  Thus, future networks will not so
much be implemented through physical changes to existing network
constructions, but will be enhanced by incorporating more intelligent
high-speed control of information flowing within the network.
Consequently, those parts responsible for providing such capabilities
must be separated from the transport layer and restructured to form an
intelligent layer.  This section first introduces the progress being
made by advanced intelligent networks that give some idea of how
advanced networks will operate.  This is followed by a description of
NTT's efforts towards software techniques that will form the core of the
intelligent layer.


5.1 Advanced Intelligent Network


In order to reach the goal of providing and popularizing personal
communications services and diverse customized services, NTT is
progressing with research and development of advanced intelligent
network services.


The basic characteristic of an advanced intelligent network is that is
architecture is even more thoroughly stratified than that of other
intelligent networks, forming a platform that does not depend on the
services provided.  Services are created and operated through the
incorporation of diverse service software within this platform.  As a
result it is not only easy to add or modify services, but it is also
possible to control complex high-level services at high speed.


Advanced intelligent network services are not only for business use;
they cover a wide-range of possibilities for home and personal use as
well.  By adding communications processing features such as voice
storage equipment, their ease of use can be greatly improved.  It is
also likely that the advanced intelligent network will grow to
accommodate even more customized services, incorporating many types of
composite services and multimedia communications.


5-2 Intelligent Software Engineering


NTT is researching and developing software production techniques from a
wide perspective, pursuing everything from supporting today's
large-scale software development program to creating a new paradigm for
the coming era of network computing.  As regards the former, we have
developed technology for preparing a computerized development support
environment for each software development process in order to produce
high-quality software effectively.  Our progress in intelligent software
engineering is an example of the latter, and is introduced below.


In software development, the fusion of two different technologies will
become very important in the future.  One is technology for acquiring
user requirements systematically and accurately from what are usually
vague descriptions.  The other is technology for recycling portions of
existing systems and applying the experience gained in earlier software
development processes.  To do this, existing systems must be analyzed
and transformed into highly abstract representations in order to
elucidate their essential functions.  It is also essential to reproduce
these functions in a suitable form for incorporation in the new system.


Assuming that software will generally be distributed over entire
networks in the future, it will become very important to develop
software that provides universally applicable functions while
simultaneously answering to a wide range of needs.  As our research into
intelligent software engineering progresses, we will continue to put
forward suggestions for new kinds of software in the age of network
computing.




6.  Basic Research


NTT's basic research is divided into six fields:  photonics, intelligent
information processing, media processing, nano-electronics,
micromachines, and the search for new materials and properties.
Pioneering results have been achieved in all of these fields.  This
section describes the latest results of our research into three of these
fields - nano-electronics, photonics, and new materials/properties --
that are predicted to revolutionize net-work technology.


6.1  Nano-Electronics -- Technology for Future LSIs


0.2mu-m LSI technology is the next step beyond 0.35mu-m, which is
currently regarded as the "next generation" of process technology, and
should become the mainstream LSI fabrication technology around the year
2000.  NTT has identified two key technologies that should bring about
substantial advances in this area, and has succeeded in manufacturing a
prototype 0.2mu-m gate array.


The first key technology is synchrotron orbital radiation (SOR), which
is electromagnetic radiation at wavelengths of 0.001mu-m, one
four-hundredth that of the ultra-violet radiation used in 0.5mu-m LSI
manufacturing.  This technology has made tremendous progress in
microprocessing possible.  The second technology, developed
independently by NTT, is the fabrication of ultra-thin silicon layers on
insulating surfaces.  By using such thin films as circuit substrates,
much higher processing speeds can be achieved.  We have successfully
developed a prototype 0.2mu-m gate array that has one-tenth the cell
area of 0.5mu-m gate arrays, and consumes ten times less power.  Using
this 0.2mu-m technology, it will be possible to implement the important
functions of today's large general-purpose computers on a single
25-mm-square LSI chip.


6.2  Photonics -- Planar Lightwave Circuits


NTT is progressing with the research and development of optical devices
and materials for the purpose of implementing an all-optical network
(including "optical" nodes) supporting photonics-through-the-network.
the main research areas here include planar lightwave circuits (PLC),
optical semiconductor devices, optical fiber amplifiers, and nonlinear
optical materials.


"Planar lightwave circuit" is a term coined by NTT to describe planar
optical circuits consisting of optical waveguides on flat substrates.
We have devised systematic and effective techniques for extracting
particular features of light waves by carefully controlling their phase
and interferences as they propagate through low-loss glass waveguides.
In other words, it has become possible to branch, couple, switch,
separate, multiplex and demultiplex optical signals without having to
transform them into electrical signals first.


PLCs are made using special fabrication techniques and advanced design
methods.  By combining the frame hydrolysis deposition techniques (as
used in manufacturing optical fibers) with LSI manufacturing processes
to produce highly advanced microprocessing techniques, we have made it
possible to produce planar waveguides of any shape.  Furthermore, by
designing and simulating circuits on computer, we are not only able to
propose new structural components, but we are also able to attain a much
higher degree of accuracy at the design stage, contributing to the rapid
creation of new products.  We are also introducing computer technology
into other areas of our research into parts and materials, as a means of
increasing the efficiency with which advances are made.


The main products to which PLC technology is applied include large-scale
star couplers, thermo-optic switches and arrayed-waveguide-grating
wavelength multiplexers.  Such devices are contributing to innovative
switching and transmission techniques.  Furthermore, we are also
researching active PLCs with optical amplification and non-linear
functions.  PLCs make it possible to achieve higher densities, greater
production volume and more diverse functions than are available with
fiber components.  It should also be possible to integrate them with
opto-electronic ICs or organic materials, and consequently PLCs should
make great contributions to the development of all-optical networks.


6.3  New Materials/Properties -- Controlled Spontaneous-Emission Diodes


> From  NTT's wide-ranging research into new materials and new properties,
our research into the field of quantum optics, which bridges the gap
between new properties and photonics, has led to a revolutionary new
device -- the controlled spontaneous- emission diode.


Spontaneous emission, the basis for light emission, is not a universal
property of atoms.  this technology is based upon the concept that, by
controlling electromagnetic fluctuations in a vacuum, we can change them
artificially.  This gave rise to the idea of producing wavelength-order
semiconductor microcavity resonators and trying to control the
electromagnetic fluctuations inside a vacuum.  In order to improve the
efficiency of photon confinement, the reflector above the microcavity is
shaped as a microlens.  This results in a laser oscillator with a
threshold that is a thousand times lower than that of a regular
semiconductor device.


The spontaneous-emission diode can output light with superior
directionality, like a laser, at close to 100% quantum efficiency even
when operating below its threshold value.  This device should thus lead
to the development of new light sources with incredible performance,
such as ultra-low-threshold semiconductor lasers and single-mode LEDs.


In addition, NTT is investigating intelligent information processing
techniques covering a wide range of fields from understanding human
perception to autonomous distributed cooperation processing techniques
that bear upon the implementation of full-scale networks.  We are also
looking into the innovation and integration of all kinds of media
technology, and we are researching micromachine technology involving the
miniaturization and fusion of advanced photonic, electronic and
mechanical elements.


7.  VI&P Experiments


NTT is progressing with comprehensive VI&P experiments by connecting its
Musashino and Yokosuka R&D Centers to produce and actual VI&P
environment.  By incorporating the latest technologies into a single
system, we hope not only to evaluate its benchmark performance under
conditions close to those of actual use, but also to add finishing
touches to the service functions.  The experiment entered its second
phase last year with the first tests of B-ISDN services, including
advanced multimedia services and high-speed computer communications.
The multimedia services are being tested using an "experimental house"
constructed near the Yokosuka R&D center.  This house is connected by an
optical subscriber communications system which supports such services as
an electronic newspaper and remote medical diagnosis by means of
super-high definition (SHD) images with a resolution similar to that of
35-mm film, and four times that of HDTV images.  The house is also
capable of simultaneously handling up to 60 high-definition TV channels
or video phone circuits.  On the other hand, our high-speed computer
communications experiments involve transmitting high-speed data (45
Mbit/s) over a distance of about 100 km from a supercomputer in
Musashino to Yokosuka via an FDDI-LAN and an ATM system.  We are also
experimenting with remote multi-party conferencing and LAN
interconnection methods.  In the future, we plan to test a wide range of
services involving, for example, applications for public networks using
ATM exchanges. [See also my report, "ntt-labs.993" 2 Nov 1993]


In the future, it is important for Japan to provide new services and
develop new software in a strategic manner, and to take as much
advantage as possible of its advanced infrastructure.  NTT's
comprehensive VI&P experiments are helping to establish how this
infrastructure should be used, and are contributing to the services it
will eventually provide.




8.  Perspective for the Future Network


Toward the next century, we anticipate a functional division of networks
between the transport layer and intelligent layer, with improved
performance for each, and the significant progress of access networks.
The network concept behind this perspective is shown in Fig. 6 [omitted,
a stylistic rendering that shows the separation between the high-speed
broadband transport layer and the intelligent layer].  Consequently,
NTT's R&D will in future place emphasis on optical frequency division
multiplexing and optical ATM technology with their potential for making
optimal use of the wide optical bandwidth, on the construction of an
access network based on new concepts, and on improvement of software
techniques as a basis for the evolution of new services.


With the ongoing enhancement and diversification of telephone services,
the fusion of information processing and communications that is expected
to take off in the lae nineties, the expected high demand for visual
services by the year 2000 and the evolution in multimedia services to
handle them, NTT is now pursuing and will continue to pursue
technological development and standardization in a wide rage of fields
on the basis of its world-wide vision and anticipation of future trends.


---------------------------END OF REPORT--------------------------