U.S. patent application number 09/766817 was filed with the patent office on 2001-10-04 for optical network.
Invention is credited to Ikoma, Yoshiaki, Sakano, Shinji, Sawada, Yasushi, Tsushima, Hideaki.
Application Number | 20010026384 09/766817 |
Document ID | / |
Family ID | 18614582 |
Filed Date | 2001-10-04 |
United States Patent
Application |
20010026384 |
Kind Code |
A1 |
Sakano, Shinji ; et
al. |
October 4, 2001 |
Optical network
Abstract
A medium-scale IP telecommunications network is configured in a
low-cost optical network with good reliability and expandability. A
physical configuration example has a center node 2-1 and eight
local nodes 2-11 through 2-18 connected in one OADM ring 2-21/2-22.
The logical configuration is a star configuration with the central
node 2-1 at its origin with all traffic passing through the center
node 2-1. The local nodes 2-11 through 2-18 are connected to the
central node 2-1 by wavelength-unit optical channels or optical
paths .lambda.1 through .lambda.8. Channels are added as required.
Initially, for example, the center node 2-1 and the local node 2-5
are connected by .lambda.5, but .lambda.13 can added when the need
arises. Since the logical star network is limited to approximately
two add/drop optical channels at local nodes, costs are reduced by
using inexpensive filters (e.g., dielectric interference film
filters or fiber Bragg reflectors) that are capable of extracting
only the specific wavelength of the optical channel. High
reliability is provided through the use of optical switches for
correction of fiber transmission path failures.
Inventors: |
Sakano, Shinji; (Kamakura,
JP) ; Sawada, Yasushi; (Yokohama, JP) ;
Tsushima, Hideaki; (Komae, JP) ; Ikoma, Yoshiaki;
(Yokohama, JP) |
Correspondence
Address: |
KNOBLE & YOSHIDA
EIGHT PENN CENTER
SUITE 1350, 1628 JOHN F KENNEDY BLVD
PHILADELPHIA
PA
19103
US
|
Family ID: |
18614582 |
Appl. No.: |
09/766817 |
Filed: |
March 14, 2001 |
Current U.S.
Class: |
398/79 ; 398/5;
398/7; 398/9 |
Current CPC
Class: |
H04J 14/0297 20130101;
H04Q 2011/0098 20130101; H04J 14/0241 20130101; H04J 14/0283
20130101; H04J 14/0206 20130101; H04J 14/0209 20130101; H04J
14/0213 20130101; H04Q 2011/0081 20130101; H04B 10/032 20130101;
H04J 14/0227 20130101; H04J 14/0219 20130101; H04Q 11/0066
20130101; H04J 14/0294 20130101; H04J 14/0238 20130101; H04J
14/0221 20130101; H04J 14/0228 20130101 |
Class at
Publication: |
359/124 ;
359/110 |
International
Class: |
H04B 010/08; H04J
014/02 |
Foreign Application Data
Date |
Code |
Application Number |
Mar 4, 2000 |
JP |
2000-100361 |
Claims
What is claimed is:
1. A wavelength division multiplexing optical network, comprising:
a client; a plurality of redundant optical paths including a
working path and a protection path; and a plurality of network
nodes redundantly connected through said optical paths in a
predetermined configuration, at least one of said network nodes
being connected to said client, said network nodes transmitting and
receiving optical signals with each other at a set of wavelengths,
each of said network nodes further comprising; an optical add/drop
multiplexer unit for converting a client optical signal at a first
channel frequency from said client to a second channel frequency at
one of said wavelengths to generate a converted client optical
signal, said add/drop multiplexer unit splitting the converted
client optical signal into split and converted redundant client
optical signals for said optical paths, said add/drop multiplexer
unit selectively multiplexing the split and converted redundant
client optical signals to generate a wavelength division
multiplexed optical signal, said add/drop multiplexer unit
selectively demultiplexing the wavelength division multiplexed
optical signal to generate a wavelength division demultiplexed
optical signal; an optical signal failure detector connected to
each of said optical paths for detecting a failure in the
wavelength division demultiplexed optical signal to generate an
optical path failure signal; and an optical switch unit connected
to said optical paths in response to the optical path failure
signal and having a first optical switch for switching from one of
said optical paths to another of said optical paths.
2. The wavelength division multiplexing optical network according
to claim 1 wherein the predetermined configuration is a ring
structure.
3. The wavelength division multiplexing optical network according
to claim 1 wherein the predetermined configuration is a liner
structure.
4. The wavelength division multiplexing optical network according
to claim 1 further comprises at least one main node and at least
two local nodes.
5. The wavelength division multiplexing optical network according
to claim 4 wherein said local nodes directly transmit the optical
signals with each other without said main node.
6. The wavelength division multiplexing optical network according
to claim 4 wherein said main node receives every one of the optical
signals at one of said wavelengths from one of said local nodes and
transmits the optical signals at a different one of said
wavelengths to another of said local nodes.
7. The wavelength division multiplexing optical network according
to claim 4 wherein said main node receives every one of the optical
signals at one of said wavelengths from one of said local nodes and
transmits optical signals at the same one of said wavelengths to
the same one of said local nodes.
8. The wavelength division multiplexing optical network according
to claim 1 wherein a number of the set of the wavelengths at each
of said network node is flexibly modified.
9. The wavelength division multiplexing optical network according
to claim 1 wherein said optical add/drop multiplexer unit further
comprises: a transponder for converting the client optical signal
at the first channel frequency from said client to the second
channel frequency at one of said wavelengths to generate the
converted client optical signal; an optical splitter connected said
transponder to for splitting the converted client optical signal
into the splitted and converted redundant client optical signals
for said optical paths; a wavelength division multiplexer connected
to each of said optical paths of said optical splitter for
selectively multiplexing the splitted and converted redundant
client optical signals to generate the wavelength division
multiplexed optical signal; and a wavelength division demultiplexer
connected to each of said optical paths of the wavelength division
multiplexing optical network for selectively demultiplexing the
wavelength division multiplexed optical signal to generate the
wavelength division demultiplexed optical signal.
10. The wavelength division multiplexing optical network according
to claim 9 wherein said optical add/drop multiplexer unit adds and
drops a single channel frequency.
11. The wavelength division multiplexing optical network according
to claim 9 wherein said optical switch unit further comprises a
second optical switch and a third optical switch connected in
series to said wavelength division demultiplexer for selectively
connecting to said first optical switch to establish a drop route
and to said wavelength division multiplexer to establish a through
route.
12. The wavelength division multiplexing optical network according
to claim 9 wherein said optical switch unit further comprises an
optical coupler connected to said wavelength division demultiplexer
for splitting the wavelength division demultiplexed optical signal
and a fourth optical switch connected to said optical coupler for
selectively connecting to said first optical switch to establish a
drop route and to said wavelength division multiplexer to establish
a through route.
13. The wavelength division multiplexing optical network according
to claim 9 further comprises an optical amplifier connected to said
wavelength division multiplexer for amplifying the wavelength
division multiplexed optical signal.
14. A wavelength division multiplexing optical network, comprising:
a client; a plurality of optical paths; and a plurality of network
nodes connected through said optical paths in a predetermined
configuration, at least one of said network nodes being connected
to said client, said network nodes transmitting and receiving a
plurality of optical signals with each other at a set of
wavelengths; a router connected to each of said network nodes for
selecting one of said optical paths for transmitting one of the
optical signals; and an optical signal failure detector connected
to each of said optical paths and said router for detecting a
failure in the optical signal to generate an optical path failure
signal for a particular one of said optical paths, wherein said
router in response to the optical path failure signal switching
from the particular one of said optical paths to another of said
optical paths for transmitting the one of the optical signals.
15. The wavelength division multiplexing optical network according
to claim 14 wherein the predetermined configuration is a ring
structure.
16. The wavelength division multiplexing optical network according
to claim 15 wherein said optical paths include dispersion shifted
fibers with their chromatic dispersion at a wavelength near 1552
nm.
17. The wavelength division multiplexing optical network according
to claim 16 wherein said dispersion shifted fibers transmit optical
input signals being multiplexed at 200 GHz interval within C-band
having a wavelength range from 1530 nm to 1560 nm, each optical
channel being less than -3.5 dBm, optical modulation speed being
2.48 Gbits/second, inter-node span loss being less than 12 dB at 40
km.
18. A method of wavelength division multiplexing for an optical
network, comprising: providing a plurality of redundant optical
paths including a working path and a protection path as well as a
plurality of network nodes redundantly connected through the
optical paths in a predetermined configuration, at least one of the
network nodes being connected to a client; transmitting and
receiving optical signals to and from the network nodes at a set of
wavelengths, said transmitting and receiving further comprising:
converting a client optical signal at a first channel frequency
from the client to a second channel frequency at one of the
wavelengths to generate a converted client optical signal;
splitting the converted client optical signal into splitted and
converted redundant client optical signals for the optical paths;
selectively multiplexing the splitted and converted redundant
client optical signals to generate a wavelength division
multiplexed optical signal; selectively demultiplexing the
wavelength division multiplexed optical signal to generate a
wavelength division demultiplexed optical signal; detecting a
failure in the wavelength division demultiplexed optical signal to
generate an optical path failure signal; and switching from one of
said optical paths to another of said optical paths in response to
the optical path failure signal.
19. The method of wavelength division multiplexing for an optical
network according to claim 18 wherein the predetermined
configuration is a ring structure.
20. The method of wavelength division multiplexing for an optical
network according to claim 19 wherein the predetermined
configuration is a liner structure.
21. The method of wavelength division multiplexing for an optical
network according to claim 18 wherein the network nodes include at
least one main node and at least two local nodes.
22. The method of wavelength division multiplexing for an optical
network according to claim 21 wherein the local nodes directly
transmit the optical signals with each other without said main
node.
23. The method of wavelength division multiplexing for an optical
network according to claim 21 wherein the main node receives every
one of the optical signals at one of the wavelengths from one of
the local nodes and transmits the optical signals at a different
one of the wavelengths to another of the local nodes.
24. The method of wavelength division multiplexing for an optical
network according to claim 21 wherein said main node receives every
one of the optical signals at one of the wavelengths from one of
the local nodes and transmits optical signals at the same one of
the wavelengths to the same one of the local nodes.
25. The method of wavelength division multiplexing for an optical
network according to claim 18 further comprising flexibly modifying
a number of the set of the wavelengths at each of the network
node.
26. The method of wavelength division multiplexing for an optical
network according to claim 18 wherein said multiplexing and said
demultiplexing add and drop a single channel frequency.
27. The method of wavelength division multiplexing for an optical
network according to claim 18 further comprising selectively
connecting to establish a drop route and to establish a through
route.
28. The method of wavelength division multiplexing for an optical
network according to claim 18 further comprising amplifying the
wavelength division multiplexed optical signal.
29. A method of wavelength division multiplexing for an optical
network, comprising: providing a plurality of optical paths and a
plurality of network nodes connected through the optical paths in a
predetermined configuration, at least one of said network nodes
being connected to a client; transmitting and receiving a plurality
of optical signals among the network nodes at a set of wavelengths;
selecting one of the optical paths for transmitting one of the
optical signals for optimal through traffic; detecting a failure in
the optical signal to generate an optical path failure signal for a
particular one of the optical paths; and switching from the
particular one of the optical paths to another of the optical paths
for transmitting the one of the optical signals in response to the
optical path failure signal.
30. The method of wavelength division multiplexing for an optical
network according to claim 29 wherein the predetermined
configuration is a ring structure.
31. The method of wavelength division multiplexing for an optical
network according to claim 29 further comprising: providing
dispersion shifted fibers with their chromatic dispersion at a
wavelength near 1552 nm in the optical paths.
32. The method of wavelength division multiplexing for an optical
network according to claim 31 wherein the optical input signals are
multiplexed at 200 GHz interval within C-band having a wavelength
range from 1530 nm to 1560 nm in the dispersion shifted fibers,
each optical channel being less than -3.5 dBm, optical modulation
speed being 2.48 Gbits/second, inter-node span loss being less than
12 dB at 40 km.
33. A ring making optical network including a dispersion shifted
optical fiber with a wave length in the vicinity of 1552 nm which
shows substantially zero chromatic dispersion on at least a part of
an optical fiber transmitting path, said optical network
comprising: a system which transmits a multiplexed optical signal
having more than two wave lengths within a wave length range of
1530-1560 nm or the C-band at an interval of 200 GHz wherein an
optical input to the dispersion shifted fiber is less than -3.5 dBm
per optical channel, a baud rate of light is less than 2.48 Gbit/s,
and a span loss between nodes is less than 12 dB, in other words,
an equivalent node interval is less than 40 Km.
Description
BACKGROUND OF THE INVENTION
[0001] The present invention is related to optical networks, and in
particular, to low-cost, general-use optical networks suitable for
transmission of traffic using Internet Protocol (IP).
[0002] Prior optical network technology was designed primarily for
voice communication. Such voice communication (telephone circuits,
etc.) employed a guaranteed service mode in which a complete
communication path between the users at both ends was guaranteed
for the duration of the call. In this guaranteed service mode,
specific users were guaranteed a specific bandwidth for the
duration of the call, regardless of actually transmitted signals.
Such guaranteed service required the construction of highly
reliable and expensive optical networks with redundant paths to
provide immediate recovery from fault conditions. Increased volume
of communications over telephone circuits has advanced speed and
capacity in optical network.
[0003] In recent years, data communication using Internet Protocol
(IP) has experienced explosive growth and has rapidly replaced
guaranteed service-type telephone circuits as the primary mode of
communication. In Internet Protocol, when a signal `packet` or a
small chunk of data arrives a router, the router routes the packet
to an open transmission path. This technique, in which a fixed
communications path need not be established between end users is
called a `connectionless` network protocol. Connectionless systems
have reduced costs because multiple users share the same signal
bandwidth. This system also features rapid recovery from faults.
When a fault occurs in a given path, after short delay to adjusted
signal flow between routers, the affect signals are routed though a
different path.
[0004] Because data communications systems using Internet Protocol
are capable of handling multimedia signals, there has been a
growing demand in data capacity not only for text, but also for
audio, graphic images, and video for personal computer
communication. Because the growing demand, greater data
transmission capacity as well as flexible expandability will be
required. Also, the areas where data is transmitted have been
expanding. Along with the need to connect more users over longer
distance, there is also a growing demand for high capacity optical
transmission of 100 Mbit/s-1 Gbit/s over distances ranging from a
few tens of kilometers to a few hundred kilometers. In the past,
optical transmission methods have primarily concentrated on
guaranteed service-type communications links as generally used for
telephone circuits to provide large capacity, high reliability, and
high quality service. Prior technology has difficulties in
providing cost reduction and flexibly responding to client
demands.
[0005] From the start, large scale optical networks have included
ring, bus, and star configurations. In general, optical network
systems had to be capable of high-speed, high-capacity,
high-reliability, and high-quality data transmission. For these
reasons, to meet the speed and capacity requirements, optical
network system designers tended to opt for techniques such as load
distribution and function distribution. Also, to meet the
reliability requirements, they used redundancy such as duplicate
circuits, hot standby circuits, etc. while to meet the quality
standards, they used QoS (quality of service) processes and TCP
(transmission control protocol).
[0006] In the past, because their designers focused primarily on
obtaining extremely high-speed, high-capacity, high-reliability and
high-quality signal transmission, optical telephone networks tended
to be too expensive for widespread use. The use of Internet
(Internet Protocol) traffic has required less speed, capacity,
reliability, and quality in internet-type communication than the
conventional voice communication circuits. A practical optical
network configuration is desired for Internet Protocol
communication in mid-sized networks such as for use in covering
metropolitan areas up to 200 km. A low-cost optical network
configuration is also desired with sufficient reliability for data
communication using Internet Protocol without abnormal congestion
under fault conditions. Also, in the past, because it was voice
communication that determined the standards for telephone circuit
networks, the required number of subscriber lines for such networks
could be accurately predicted from the number of residences and
offices in a given service area. However, because data networks
handle everything from simple text to high quality video, networks
now must have sufficient flexibility for expandability so as to
accommodate a broad range of traffic volume at low cost. It is
desired that optical networks have low initial costs, but are
easily upgradeable for an operator to have a good cash flow.
[0007] In consideration of the above described issues, it is an
object of the present invention to provide a low-cost optical
network configuration for use primarily in medium-scale IP
networks. It is a further object of the present invention to
provide optical networks that will
[0008] provide the client with the kind of signal capacity demanded
by fast-changing IP networks.
[0009] provide the flexibility to support communication formats
such as SONET and Ethernet
[0010] require a lower initial capital investment
[0011] provide sufficient expandability for growth, and
[0012] provide steady cash flow.
[0013] It is a further object of the present invention to reduce
total costs by making effective use of the limited bandwidth
resources of optical networks. For optical transmission over
distances in excess of 40 km, optical repeaters such as optical
fiber amplifiers are required. In recent years, the use of EDFA
(Erbium-Doped Fiber Amplifiers, that is made from silica fibers
doped with erbium) as optical amplifiers has become commonplace.
These optical amplifiers for a wavelength band range of 1530-1560
nm (C-band) have simple components, while it is technically
feasible to build EDFAs that will amplify across the entire L-band
(1570-1610 nm), the optical amplifiers for this range have complex
components. In one possible optical network configuration for that
wavelength band range, unless high-precision and expensive
components known as wavelength lockers are used, the minimum
wavelength separation that can be obtained is around 200 GHz.
Therefore, in terms of a power of 2 the number of possible C-band
or L-band wavelengths is about 16 channels or wavelengths at most.
A way to make effective use of this limited [bandwidth] resource is
desired.
[0014] It is also an object of the present invention to provide
optical networks in which common components are used regardless of
the types of signals connectable to the client, the types of signal
transmission paths, or the usable bandwidth. Dispersion-shifted
optical fibers are used in a transmission path to shift the
zero-dispersion wavelength to near 1552 nm, where less expensive
C-band optical components can be used. However, when a C-band
wavelength division multiplex signal with equal wavelength spacing
between channels is transmitted over 40 km in such a path at the
normal optical transmission path levels recommended in ITU-T, a
phenomenon known as `four wave mixing` (an interference mode
between two wavelengths) can occur. Four wave mixing causes
overlapping of equally spaced signals, which degrades the optical
transmission characteristics of the path. The problem, then, is to
devise a way of using low-cost C-band optical components for
wavelength-division multiplex transmission in dispersion-shifted
optical fiber. One technology already established for eliminating
the effects of the four wave mixing phenomenon is `unequal
spacing.` Because this technology requires specially designed
optical components, it is not conducive to cost reduction. These
are the issues that must be improved.
SUMMARY OF THE INVENTION
[0015] Following are some features of the present invention:
[0016] Optical switches are used for `ring protection` in the IP
optical transmission mode. This provides a flexible protection
scheme that is independent of the service type or signal format of
the client.
[0017] Connection management is made easier by the use of optical
add/drop multiplexers in physical ring/logical star
configurations.
[0018] Cost reduction is realized through the use of single
wavelength insertion/extraction filters for specific wavelengths at
local nodes.
[0019] Provides expandability through addition of single wavelength
insertion/extraction filters for specific wavelengths at local
nodes.
[0020] Fixed-connection channels and flexibly recombinable channels
provide flexibility in additional channel assignment.
[0021] Dynamic optical switching functions are incorporated in
additional channels.
[0022] Mesh connections are incorporated in additional
channels.
[0023] Optical multicast functions are incorporated in additional
channels
[0024] A `traffic healing` configuration for multiple path
connections provides high through-put during normal conditions and
traffic healing during fault conditions.
[0025] The use of C-band (1530-1560 nm) in dispersion-shifted
optical fiber allows the use of low-cost optical components.
[0026] In incorporating optical multicast and direct optical
switching, optical amplifiers are added as required to eliminate
drop signal loss and changes in optical level when optical paths
are switched, thus avoiding variations in optical output.
[0027] According to a first means of the present invention for
solving the above problems, in a central node or local node of an
optical network having a certain physical ring configuration, with
two optical paths for propagating optical signals in a clockwise
and counterclockwise directions; at the central node or local node,
in a first direction, a signal that constitutes the optical signal
from the router side within the node the first means includes:
[0028] a transponder for performing wavelength conversion to a
wavelength corresponding to one of the channels of a
wavelength-division multiplexed signal that is propagated though a
transmission path of the optical fiber ring;
[0029] an optical divider for splitting the optical signal from the
transponder into clockwise- and counterclockwise-propagating
signals;
[0030] an optical insertion unit further comprising
wavelength-division multiplexers installed for the two paths, for
multiplexing optical signals of various wavelengths; a wavelength
band optical insertion unit installed for the two paths, for adding
L-band channels as required; an optical insertion unit capable of
inserting an optical signal to be used for monitor and control
(Optical Supervisory Channel);
[0031] wherein a second direction signal from another node that is
passed through the two optical fiber transmission paths of the ring
and input to that node is input, and, installed for each of the two
paths, as required, are
[0032] an `Optical Supervisory Channel` optical signal extraction
unit;
[0033] an optical extraction unit with extraction means capable of
extracting light of the wavelength band used for additional L-band
channels;
[0034] an optical switch for selecting the output in the second
direction from the two paths during normal non-fault conditions and
the other signal in the other direction when a fault condition is
detected;
[0035] a detector for detecting fault conditions; and
[0036] a transponder for converting optical signals from the
optical switch to an optical signal of a wavelength that is
received by the router in that node.
[0037] The network is configured by the optical divider and the
optical switch such that optical signals in the ring path are split
into the two signals for the transmit side and the receive side,
with redundant systems provided for propagation of the signals in
the two paths. Also, because in the center node, all of the optical
channels are extracted and wave-mixed, there may be different
losses resulting in level variations between the optical power
levels of the various channels. Therefore, in order to equalize
these levels, optical power level controllers for each of the two
paths are placed before the wavelength-division multiplexers for
adjusting the optical power levels of the optical signals extracted
in the two paths to substantially the same optical power
levels.
[0038] In telephone networks, when an optical transmitter or
receiver in a transponder fails, operation must be immediately
restored (e.g., within 50 ms), whereas IP communication systems can
tolerate fault recovery for a few minutes or even up to a few
hours. The better time allows to replace a card if an on-site spare
is available. This means that to the extent that problems can be
corrected by replacing parts, equipment need not necessarily have
protection functions provided by the redundant circuits. A failure
in a fiber transmission path, however, is not something that is
easy to fix in a few minutes or hours. Therefore, to prevent from
affecting IP traffic, systems must have redundant circuits to
respond to breaks in optical fibers. Thus, in the IP communications
system of the present invention, the two paths in two-path ring
configurations are used not only for transmitting and receiving,
but also to provide fault recovery protection in conjunction with
an optical switch. This provides significant cost reduction. Also,
in IP communication, there typically are large fluctuations in
demand at the client router optical interface, and to accommodate
these fluctuations, the flexibility is required to handle a variety
of signal types and protocols. In the present invention, this
requirement is satisfied by using optical switches whose operation
does not depend on signal type.
[0039] According to a second means of the present invention for
solving the above and other problems is an optical network that has
a two-fiber transmission path ring as its physical configuration,
but has a star logical configuration wherein all traffic passes
through the central node with respect to the wavelength units that
are the signal connection units for optical signals between the
central and local nodes. For example, when 16 C-band wavelengths
are used, the use of a star logical configuration makes it possible
to connect signals among up to 15 local nodes by way of the central
node. This provides a network configuration that makes effective
use of the limited bandwidth resource. In a full mesh connection,
16 wavelengths can support connections among no more than 4 nodes.
Also, since wavelength-division multiplex configurations with
point-to-point connections require multiple repeater nodes between
widely spaced nodes, communication costs are high. In the logical
star configuration, however, the center node is the only repeater
node, the above wasteful cost is substantially reduced.
[0040] According to a third means of the present invention for
solving the above problems, an optical network having a central
node and a local node in an optical fiber ring configuration as its
physical configuration, the local node comprises, for a
wavelength-multiplexed optical signal input from an optical fiber
transmission path arranged in a two-ring configuration
[0041] an insertion/extraction unit for extracting a wavelength
that corresponds to one of the channels in the
wavelength-multiplexed optical signal, inserting an optical signal
from a router in the local node as an optical signal that has been
converted to the extracted wavelength, and outputting a
wavelength-division multiplexed optical signal;
[0042] a node output structure that is added as required and has an
optical insertion unit which comprises:
[0043] an insertion means to which the
wavelength-division-multiplexed
[0044] optical signal from the above insertion/extraction unit is
input the
[0045] insertion unit being capable of inserting
[0046] channel-addition-wavelength-band light, and
[0047] an optical insertion means capable of inserting an optical
supervisory channel signal;
[0048] an optical extraction unit for outputting a
wavelength-division multiplex signal to the above
insertion/extraction unit comprising, as required,
[0049] a first extraction unit, to which is input a
second-direction signal which,
[0050] along with the multiplexed light signal, is output to a
router within the node,
[0051] the extraction unit being capable of extracting a
supervisory channel light signal, and
[0052] a second extraction unit to which can be added extraction
unit capable of
[0053] extracting channel-addition-wavelength-band light;
[0054] a transponder for inputting a light signal from a router in
the node in the first direction optical to the central node, and
converting it to a wavelength corresponding to one of the channels
of a wavelength-division multiplex signal, and also converting an
optical signal to output in a second direction from a wavelength
corresponding to one of the channels of the wavelength-division
signal to a wavelength that is received by the router; and
[0055] an optical switch means comprising
[0056] an optical switch which, for a first direction, splits the
optical signal from the transponder into two signals in order to
output it to the two ring paths in the clockwise and counter
clockwise directions, and output to the above two
insertion/extraction units; and, in a second direction, selects the
optical signals from the above two insertion/extraction units that
were input along the two paths, and outputs them to the above
transponder; and
[0057] a detector for detecting failure in the signal from the
above wavelength extraction unit;
[0058] wherein as a logical configuration, a wavelength-unit star
configuration is provided for the connection of signals between the
local nodes and a single central node such that all traffic passes
through the central node.
[0059] According to a fourth means of the present invention for
solving the above problems, is an optical network comprising a
central node and local nodes of the optical network made such that
the insertion/extraction of the optical signals of a plurality of
channels is accomplished by connecting, in the configuration of a
local node, a plurality of insertion/extraction means such that a
wavelength-multiplex mixed-wave output from a first
insertion/extraction means is output to a wavelength-multiplex
mixed-wave input of a second insertion/extraction means, and the
wavelength-multiplex mixed-wave output from the second
insertion/extraction means is input to the wavelength-multiplex
mixed-wave input of a third insertion/extraction means. Because
additional optical channels are established in local nodes and IP
communications networks based on a logical star configuration, the
transmission capacity thereof is freely expanded in response to
demand. In this method, because only the required number of
wavelengths is inserted and extracted, when the number of
inserted/extracted optical channels is small, the cost is low in
comparison to other methods in which the wavelength extraction is
performed for all of the wavelengths. For these reasons, a steady
cash flow is maintained.
[0060] According to a fifth means of the present invention for
solving the above and other problems, optical switching means are
add/drop/through optical switches for selecting either a drop route
for extracting a signal to be output in a second direction, or a
through route for looping-back the signal to the ring. In this
configuration, the channel connections via the switching speed of
the add/drop/through optical switches can easily be changed in a
matter of milliseconds by remote control, via the network operation
system. In the logical star configuration, communication between
nodes always has to pass through a center node router in this
system. The connections can be made optically to reduce the number
of routers through which the signal must pass, thus reducing
cost.
[0061] Although the system above described employs dynamic optical
switching and remote control via the operation system of the
network to make direct local-node-to-local-node optical channel
connections, costs could be reduced by employing manual switching
of optical connectors to create a partial mesh configuration to
thus reduce the traffic passing through central node routers.
[0062] According to a sixth means of the present invention for
solving the above problems, the above optical switch further
comprises an optical coupler for splitting a signal to be output in
the second direction into a drop route for dropping the signal or a
through route to loop it back. In this configuration, because the
same signal is received as an optical signal by multiple nodes, the
present invention provides more extensive broadcasting capability
than a router multicast function and is less expensive than
electrically switched router multicasting in a large transmission
capacity.
[0063] According to a seventh means of the present invention for
solving the above problems, additional optical amplifiers are added
to the configuration to compensate for losses occurring in optical
switches, etc., that are in the pass-through state in dynamic
optical switching and optical multicasting, and optical loss
occurring at optical multicast optical drops. The present means
would be needed in optical networks to provide better service than
that provided by router functions for IP communications.
[0064] According to an eighth means of the present invention for
solving the above problems, in a dual ring configuration, with
respect to the flow of signals from central nodes to local nodes,
for the signal of one of the add channel wavelengths in the optical
signals transmitted in each of the two optical fibers (clockwise
and counterclockwise), the optical signal from an independent
router output is converted by an independent transponder to the
same wavelength as that of a vacant optical channel so that the two
independent optical signals transmitted at the same wavelength. In
the local node at the receive end, the independent signals from the
two optical fiber transmission paths in clockwise and
counterclockwise directions are received by a transponder in which
the light of the two signals is kept independent, and is connected
to independent routers. During normal (non-fault) conditions the
flow between routers is adjusted for maximum throughput. When a
failure occurs in one of the fiber paths, the loss of signal is
detected by the optical receiver and adjustments are made between
routers to overlay signals in good paths to restore the
operation.
[0065] According to a ninth means of the present invention for
solving the above problems, an optical ring system is provided in a
dual optical fiber transmission path ring configuration, using the
wavelength near 1552 nm (dispersion-shifted fiber) at which the
wavelength dispersion will be zero in at least one portion of the
optical fiber transmission path. In a system for
multiplex-transmission of at least two wavelengths in the 1530-1560
nm wavelength range (C-band) at a spacing of 200 GHz, in which the
optical level input to the dispersion-shifted fiber is limited to a
maximum of -3.5 dBm per optical channel, and the modulation rate is
limited to a maximum of 2.48 Gbit/s, with a maximum inter-node span
loss of 12 dB (for an equivalent node spacing of 40 km). In this
system, under the worst case, zero dispersion wavelength conditions
with a four-wave mixing interference component as a disturbance
affecting other channels (degrading the bit error rate worse than
10E-12), the system optical channel unit output level is above -3.5
dBm. Thus if the optical output is kept below -3.5 dBm, the
influence of four-wave mixing disappears. Because the prior
research and development were centered around rates of at least 10
Gbit/s using conventional dispersion-shifted fiber, at an optical
output of -3.5 dBm it was not possible to come up with a power
level diagram [(a power budget)] that would provide an adequate S/N
(signal-to-noise ratio) in a high optical power level system
linking fifteen or so repeater nodes with a required minimum
received light level of -16 dBm. On the other hand, for a Gigabit
Ethernet with a modulation rate of 1.25 Gbit/s or a 2.48 Gbit/s
STM16 modulation rate, the minimum optical power level required for
a high S/N ratio is set as much as 9 dB or 6 dB lower than that
required for a 10 Gbit/s rate, and a 16-node ring network
configuration is possible. In particular, because of the fact that
C-band optical components are used, not only inexpensive optical
components designed for standard dispersion-shifted fiber
transmission paths are used, but also channels are added to as many
as 32 channels spaced at 200 GHz intervals (including L-band) by
using an optical network having a dispersion-shifted fiber
transmission path. Thus the current invention provides flexible
expandability.
[0066] According to the tenth method of the present invention, a
ring making optical network of two fiber transmitting paths
including a optical fiber, in other words, a dispersion shifted
fiber with a wave length in the vicinity of 1552 nm shows no
chromatic dispersion on at least a part of an optical fiber
transmitting path. The above mentioned optical network includes a
system which transmits a multiplex optical signal having more than
two wave lengths within a wave length range of 1530 nm-1560 nm, in
other words, the C-band at an interval of 200 GHz. In the above
mentioned optical ring network, an optical input to the dispersion
shifted fiber is less than -3.5 dBm per optical channel, an optical
baud rate is less than 2.48 Gbit/s, and a span loss between nodes
is less than 12 dB, in other words, an equivalent node interval is
less than 40 km. Under the worst condition that the chromatic
dispersion is 0, when an optical input is -3.5 dBm or more than
-3.5 dBm per optical channel, an interfering constituent generated
by the four wave mixing disturbs and influences the wave length of
another channel, in other words, the bit error rate gets worse than
10E-12. Therefore, when an optical input is less than -3.5 dBm per
optical channel, the above mentioned influence is removed.
[0067] So far, a dispersion shifted optical fiber with a rate of
more than 10 Gbit/s has been mainly researched and developed.
Therefore, an optical input of -3.5 dBm per optical channel is too
high to give a sufficient S/N (signal to noise ratio), and the
system relayed at a plural node, for example, 15 nodes needs a
reception optical level of more than -16 dBm, resulting in the
failure to construct a level diagram. On the other hand, Gigabit
Ethernet with a baud rate of 1.25 Gbit/s or STM16 with a baud rate
of 2.48 Gbit/s has the lowest optical level to give a high S/N 9 dB
or 6 gB lower than the above mentioned optical fiber with a baud
rate of 10 Gbit/s, respectively, resulting in success of
constructing a ring of 16 nodes. In particular, since an optical
component for the C-band is used, inexpensive optical component is
for a transmitting path of conventional dispersion shifted optical
fiber is used. In addition, an optical network including a
transmitting path of a dispersion shifted optical fiber, which has
32 or less than 32 channels including an L-band at intervals of 200
GHz is constructed, resulting in an improvement of the
extensibility of an optical network.
BRIEF DESCRIPTION OF THE DRAWINGS
[0068] FIGS. 1A and 2B illustrates diagrams illustrating physical
and logical connections two-fiber UPSR (unidirectional path
switching ring).
[0069] FIGS.2A and 2B illustrates diagrams illustrating physical
and logical connections another two-fiber UPSR.
[0070] FIGS. 3A and 3B illustrates diagrams illustrating a
two-fiber UPSR with protection path from central node 1 to local
nodes.
[0071] FIG. 4 illustrates a diagram illustrating a protectionless
path configuration.
[0072] FIG. 5 illustrates a diagram illustrating a mixed logical
star/logical mesh configuration.
[0073] FIG. 6 illustrates a diagram illustrating a MPLS
(multiprotocol labeled switching) support through direct optical
switching.
[0074] FIG. 7 illustrates a diagram illustrating a multicast
support configuration.
[0075] FIGS. 8A and 8B are diagrams illustrating traffic healing
configuration with a failure in one of its two independent signal
paths.
[0076] FIG. 9 illustrates a diagram illustrating a linear OADM
network having linearly arranged geographic node locations.
[0077] FIG. 10 is a diagram illustrating a main signal flow in
center node of two-fiber UPSR network having optical add/drop
multiplexers in a logical star configuration.
[0078] FIG. 11 is a diagram illustrating a rack configuration
(1).
[0079] FIG. 12 is a diagram illustrating a main signal flow in
local node of two-fiber UPSR network having optical add/drop
multiplexers in a logical star configuration (1).
[0080] FIG. 13 is a diagram illustrating a main signal flow in
local node of two-fiber UPSR network having optical add/drop
multiplexers in a logical star configuration (2) FIG. 14 is a
diagram illustrating a main signal flow in local node of two-fiber
UPSR network having optical add/drop multiplexers in a logical star
configuration (1) FIG. 15 is a diagram illustrating a main signal
flow in local node of two-fiber UPSR network having optical
add/drop multiplexers in a logical star configuration (2).
[0081] FIG. 16 is a diagram illustrating a rack configuration
(2).
[0082] FIG. 17 is a diagram illustrating a main signal flow in
local node of two-fiber UPSR network having optical add/drop
multiplexers in a logical star configuration (1).
[0083] FIG. 18 is a diagram illustrating a main signal flow in
local node of two-fiber UPSR network having optical add/drop
multiplexers in a logical star configuration (2).
[0084] FIG. 19 is a diagram illustrating a main signal flow in main
node of two-fiber UPSR network having optical add/drop multiplexers
(1).
[0085] FIG. 20 is a diagram illustrating a main signal flow in main
node of two-fiber UPSR network having optical add/drop multiplexers
(2).
[0086] FIG. 21 is a diagram illustrating a rack configuration for
main node with add/drop optical switches.
[0087] FIG. 22 is a diagram illustrating a rack configuration for
local node with add/drop optical switches.
[0088] FIG. 23 is a diagram illustrating a main signal flow in main
node of two-fiber UPSR network having optical add/drop multiplexers
(1).
[0089] FIG. 24 is a diagram illustrating a main signal flow in main
node of two-fiber UPSR network having optical add/drop multiplexers
(2).
[0090] FIG. 25 is a diagram illustrating the C-band multiplex
wavelength transmitting system.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
1. Optical Network Configurations
(1) Optical Network Example 1
[0091] A two-fiber UPSR (unidirectional path switched ring) is
configured as the ring network. Here, the term `path,` is
synonymous with optical channel. A diagram of the two-fiber UPSR is
shown in FIG. 1. The two-fiber UPSR has one fiber ring 2-21 for the
clockwise direction, and another fiber ring 2-22 for the
counterclockwise direction. One fiber is used as a working path,
while the other as a protection path. In a typical configuration,
one center node 2-1 and a maximum of eight local nodes 2-1 1
through 2-18 are connected in OADM (optical add/drop multiplexer)
rings 2-21, 2-22. In the present example, FIG. 1A shows the
physical configuration that includes one central node 2-1 and eight
local nodes 2-11 through 2-18. The logical configuration of the
network is shown in FIG. 1B. In this example, the logical
configuration is a star configuration with the central node as the
common node through which all traffic must pass.
[0092] For the initial wavelength assignments between the central
and local nodes, the local nodes 2-11 through 2-18 are connected to
the central node 2-1 by the wavelength unit channels (optical
paths) .lambda.1 through .lambda.8 (indicated by solid lines in the
drawing). To add channels (e.g., to +1 wavelength), the wavelengths
.lambda.9 through .lambda.16 (dotted lines) can be used to add
optical channels as required. Initially, for example, the center
node 2-1 and the local node 2-5 are connected by .lambda.5, but
.lambda.13 can be added if required. The same applies to the other
nodes. The logical star network provides efficient utilization of
the limited optical channel resources. Limiting the add/drop
optical channels to about two channels per logical star local node
makes it possible to use inexpensive interference dielectric film
filters and fiber Bragg reflectors to extract only a
specific-wavelength channel. Thus the above described
implementation reduces the costs.
[0093] FIG. 2 shows another two-fiber UPSR configuration. The
invention is not confined to application in the usual geographical
urban configuration, in which the center node 2-1 is the sole large
communication center with relative small nodes. FIG. 2 shows a
configuration, for example that includes large local nodes 2-19 and
2-20 that can accommodate as many as 3-8 channels (as in the
physical and logical configurations as respectively shown in FIG.
2A and 2B). This example also uses a star configuration with a
common central node (origin) though which all traffic must pass.
Wavelength unit channels (optical paths) .lambda.1 through
.lambda.8 (solid lines) are allocated for the initial wavelength
assignments between the central and local nodes. To add channels to
a large local node (e.g., up to +7), wavelengths up to .lambda.16
(dotted lines) are used.
[0094] In addition to the fixed channel .lambda.3, the added
channels .lambda.10 through .lambda.12 (dotted lines) are connected
between the central node 2-1 and the large local node 2-19.
Channels .lambda.14 through .lambda.16 can be connected in addition
to the fixed channel .lambda.6 between the central node 2-1 and the
large local node 2-20. Thus the maximum number of channels at a
local node is selectable for improving its cost effectiveness over
nodes having only two channels. By configuring additional large
node combinations, ring networks have added flexibility at a lower
total cost.
(2) Optical Network Example 2
[0095] FIGS. 3A and 3B show a two-fiber UPSR configuration that has
a protection path from the central node 2-1 to a local node 2-12.
The two-fiber UPSR has one fiber 2-21 for the clockwise direction
and the other fiber 2-22 for the counterclockwise direction. One
fiber being used as a working path as shown in FIG. 3A while the
other as a protection path as shown in FIG. 3B. If a failure occurs
in the working path, the path is switched so that the signal can be
received via the protection path that is the opposite direction
path as shown in FIG. 3B. The system thus provides fault recovery
for signals in the faulty span. In this example, an IP signal from
a first router 2-30 is converted to a specific wavelength by a
first transponder 2-31, and the signal is divided to travel the
clockwise path 2-21 and counterclockwise path 2-22 by an optical
divider 2-32. In the receive node 2-12, the optical signal is
selected from the clockwise path 2-21 or counterclockwise path 2-22
by an optical switch 2-33, from which it is connected to a second
transponder 2-36. The second transponder 2-36 converts the signal
to the receive wavelength of a second router 2-37 and sends the
output to the router 2-37. When an optical fiber path failure 2-35
occurs, the receive node 2-12 detects an optical input interruption
and switches the optical switch 2-33 to receive the optical signal
from the protection path 2-21. After a specific time period
referred to as (1+1), the receive node 2-12 performs a
non-switch-back path-switching operation.
(3) Optical Network Example 3
[0096] FIG. 4 shows a protectionless path network configuration. In
this configuration, separate optical channels having the same
wavelength are established. For example, in the connection from the
central node 2-1 to the local node 2-12, or from the local node
2-12 to the central node 2-1, one data traffic travels in the fiber
route 2-38 and the other in the fiber route 2-39. Thus, a double
UPSR path is established. However, when there is a fault in either
of the transmission route, no quick recovery is made due to the
lack of protection.
(4) Optical Network Example 4
[0097] FIG. 5 shows a combined logical-star/logical-mesh network
configuration. Basically, an optical channel must be established
between the central node 2-1 and each of the local nodes 2-11
through 2-18 in the logical star configuration This forces even a
small amount of local-node-to-local-node signal traffic to be sent
through the central node 2-1. That is, when heavy traffic occurs
between pairs of local nodes such as between 2-11 and 2-14, 2-12
and 2-17, and 2-15 and 2-16 the configuration still routes
everything through the central node 2-1 and optical channels of
different wavelengths are required between the central node 2-1 and
each of the local nodes. This uses up a large number of wavelength
channels and also creates the need for a separate router at the
central node 2-1 for signals that simply pass through it. By
providing single direct optical channels between selected pairs of
local nodes such as channels .lambda.9, .lambda.11, and .lambda.12
between the local nodes 2-11 and 2-14, 2-12 and 2-17, and 2-15 and
2-16 as shown in dotted lines, we can avoid this wasteful use of
channels and the expense of the additional router. This is
accomplished by giving the central node 2-1 the capability to pass
optical channels through it without going through channel
terminations or optical transceivers. The capability to add
wavelength channels to the local nodes 2-11 through 2-18 as desired
is thus provided.
(5) Optical Network Example 5
[0098] FIG. 6 shows a direct optical switching network to support
MPLS (multi-protocol labeled switching) traffic. The direct optical
switching network provides the capability to support MPLS, which is
required in Internet Protocol. To increase scalability, it uses
optical switches to automatically establish optical channels by
remote control. Basically, an optical channel must be established
between the center node 2-1 and each of the local nodes 2-11
through 2-18 in the logical star configuration. In particular, a
switching function with approximately 1 ms switching time is
provided to add channels between the center node 2-1 and two local
nodes such as 2-19 and 2-20 as required to provide enough bandwidth
to pass high volume MPLS-compliant IP signals. By establishing
multiple wavelengths (e.g., .lambda.9 through .lambda.16) directly
between three nodes (e.g. 2-1, 2-19, and 2-20) rather than by
providing the required MPLS node-to-node capacity through various
routings, the current invention reduces the number of routers
required just for the routing of communications and the level of
control required to enhance flexibility. This approach provides the
high-capacity MPLS support by providing dynamic switching functions
at specific nodes thus makes efficient use of the finite number of
wavelengths such as 16 wavelengths available in the ring network.
Changes to direct switching PCBs at the large nodes and
augmentation of some control functions are performed as need
dictates. Also, because signals will be degraded by losses incurred
when they pass through wavelength add/drop and optical switch
devices for dynamic switching. Optical amplifiers are installed to
compensate for this dynamic switching loss.
(6) Optical Network Example 6
[0099] FIG. 7 shows a configuration to support optical
multicasting. IP routers have multicast functions, but delays by
each extract/repeat process add up to the point where the delay
time causes problems for some applications. High broadcast
capability or capability to transmit simultaneously to a large
number of nodes demands that the direct extract/repeat process be
performed optically. Basically, an optical channel must be
established between a center node 2-1 and each of the local nodes
2-11 through 2-18 in the logical star configuration. To distribute
a signal from the center node 2-1 at which high broadcast
capability is required, for example, the extract/repeat function
would have to be performed at each of the local nodes 2-11, 2-12,
2-14,2-15, 2-17, and 2-18 that want to receive the signal in
optical channel .lambda.9. The last local node 2-18 may serve the
termination point, but the termination node may also serve as a
center node 2-1 so that distribution of signals is confirmed.
Changing of multicast PCBs and augmentation of some control
functions are performed as the need dictates. Because the
drop/through process is optically performed on the multicast signal
channel .lambda.9 at each of the local nodes 2-11, 2-12, 2-14,
2-15, 2-17, and 2-18, it results in an optical power drop loss that
degrades the signal. Therefore, optical amplifiers are provided at
local nodes to compensate for the drop loss.
(7) Optical Network Example 7
[0100] FIGS. 8A and 8B show a traffic healing configuration that
enables operation to be restored when independent optical signals
are connected in two different paths and a failure occurs in one of
the paths. Rather than obtaining fault protection by using the
optical divider 2-32 and optical switch 2-33, as described above in
FIG. 3, between a transmit and receive nodes 2-1 and 2-12, the OADM
system 2-34 of this configuration has two transponders 2-31A and
2-31B for two optical channels from a send router 2-30. The same
two signals from two transponders 2-31A and 2-31B are routed
through two independent paths 2-21 and 2-22 to two transponders
2-36A and 2-36B provided at the receive node 2-12, where they are
received and processed by a router 2-37. In the normal or non-fault
state, the independently transmitted signals provide high
throughput as shown in FIG. 8A. When a failure occurs in one of the
paths 2-35, loss of optical signal (LOS) is detected at the receive
node 2-12 as shown in FIG. 8B. The LOS detect signal results in a
fault alarm to the transmit node 2-1, which diverts the signals
from the faulty side and overlays them in the good path 2-21 to
restore service.
(8) Optical Network Example 8
[0101] FIG. 9 shows a linear OADM network featuring a geographic
configuration in which the nodes are arranged in a linear fashion.
The physical configuration includes rightward fiber paths 2-61A and
2-61B as well as leftward fiber paths 2-62A and 2-62B. Unlike ring
configurations, this configuration provides no protect function
using an optical divider 2-32 and optical switch 2-33, as described
above with respect to FIG. 3. A logical star configuration for this
network is implemented substantially the same as in Examples, 1, 4,
5, 6, and 7.
2. Nodes
(1) Node Example 1
[0102] FIG. 10 shows the main signal flow in an OADM (optical
add/drop multiplex) central node in a logical star two-fiber UPSR
(unidirectional path switched ring) optical network.
[0103] The OADM system receives its main signal input from an
external Gigabit Ethernet (GbE) interface unit. In the OADM system,
the main signal or the add side signal in the first direction from
the GbE unit 3 -1 is input to a transponder 3-2, that
wavelength-convert the input. The wavelength conversion process in
transponder 3-2 temporarily converts the incoming signal from an
optical signal to an electrical signal by an O/E converter. An E/O
converter then converts this electrical signal to an optical signal
having the wavelength of one of the channels of the wave-division
multiplexed signal of the ring and outputs it. The signal thus
output from the transponder 3-2 is split in two by an optical
divider in the switch unit 3-3 and sent respectively to the
automatic optical power level controllers 3-4 and 3-5 or ALC (0)
and ALC (1). In the automatic optical power level controllers 3-4
and 3-5, the two signals are adjusted to a specific optical power
level and sent respectively to the wavelength-division multiplexers
3-6 and 3-8 or MUX (0) and MUX (1). In the wavelength-division
multiplexers 3-6 and 3-8, the add input wavelengths are multiplexed
and transmitted. In the L-band optical transmitter/amplifiers 3-10
and 3-11 or OTA (0) and OTA (1), these multiplexed optical signals
are amplified by an optical amplifier for transmission in the
transmission path. Each of the L-band optical
transmitter/amplifiers 3-10 and 3-11 has an optical insertion unit
for inserting add-channel-wavelength-band optical signals and an
optical insertion unit for inserting an OSC (optical supervisory
channel) optical signal. The add-channel-wavelength-band optical
signals are the band used for adding channels, or L-band in the
example shown in FIG. 10.
[0104] The main signals in the second direction or the drop-side
signals are received from the OADM units 3-14 and 3-15, which are
respectively the nodes next to the present node in the 0-path and
1-path fiber rings. The drop-side signals are input respectively to
the 0-path and 1-path input bandpass filters 3-12 and 3-13 or
BPF-IN (0) and BPF-IN (1). Each of the input bandpass filters 3-12
and 3-13 has optical extraction means for extracting OSC signals
and an optical extraction means for extracting
add-channel-wavelength-band (L-band in the drawing) signals. In the
wavelength-division demultiplexers 3-7 and 3-9 or DMUX (0) and DMUX
(1), the incoming multiplexed signals are separated into their
individual wavelength channel signals, which are fed to a (fixed
channel switch) optical switch 3-3. Provided in the optical switch
3-3 is a switch SW for selecting signals input from the
wave-division demultiplexers 3-7 and 3-9. Provided at the optical
input terminations of the optical switch 3-3 whose inputs are from
the demultiplexers 3-7 and 3-9, are LOS (loss of optical signal)
detectors 3-40 and 3-41, for detecting faults and performing
optical switching as required to restore service.
[0105] A rack-mounting configuration is shown in FIG. 11. The
equipment required to be installed initially includes a transponder
rack 3-50, an optical switch and wavelength-division
multiplexer/demultiplexer rack 3-51, and an optical
transmitter/amplifier rack 3-52. The units that are required
initially to support eight channels include transponders 3-2-1
through 3-2-8, optical switches 3-3-1 though 3-3-8, `0` and `1`
automatic optical power level controllers 3-4 and 3-5, `0` and `1`
wavelength-division multiplexers 3-6 and 3-8, `0` and `1`
wavelength-division demultiplexers 3-7 and 3-9, `0` and `1` L-band
optical transmitter/amplifiers 3-10 and 3-11, and an OSC (optical
supervisory channel) signal processor 3-60. Each of the units 3-2-1
through 3-2-8 and 3-3-1 through 3-3-8 supports one of eight
different-wavelength channels. In addition, when it is desired to
add channels, an additional transponder rack 3-50 and an optical
channel wavelength multiplexer rack 3-51. By standardizing the
add-on unit configuration, the number of different kinds of
equipment is reduced, and that in turn reduces cost.
(2) Node Example 2
[0106] FIGS. 12 and 13 respectively show main signal flows (1) and
(2) in an OADM local node in a logical star two-fiber UPSR optical
network.
[0107] On the add side, the main signal from an external Gigabit
Ethernet (GbE) interface unit 3-1 is fed to a transponder 3-2 in
the OADM system. Transponder 3-2 performs wavelength conversion. In
an optical switch 3-3, the OADM output is divided into a 0-path
signal and a 1-path signal as an input to the channel add/drop
units 3-16 and 3-17 or ChADM1 (0) and ChADMI (1). In the add/drop
units 3-16 and 3-17, the signals from the optical switch 3-3 are
multiplexed with other channels and sent on to the L-band optical
transmitter/amplifiers 3-10 and 3-11 or OTA (0) and OTA (1). If
there is only a short distance between the present node and the
OADM units 3-14 and 3-15 which are the nodes next to the present
node in the 0-path and 1-path fiber rings, and the loss between
these nodes is small, no optical amplifiers will be required at
this point. In this case, the L-band optical transmitters 3-18 and
3-19 of FIG. 13, which have no amplifiers, are alternatively
replaced with the L-band optical transmitter/amplifiers 3-10 and
3-11 of FIG. 12.
[0108] Still referring to FIG. 12, the main signals in the output
direction (the drop-side signals) are received from the OADM units
3-14 and 3-15, which are respectively the nodes next to the present
node in the 0-path and 1-path fiber rings, and are input
respectively to the input bandpass filters 3-12 and 3-13 or BPF-IN
(0) and BPF-IN (1). In each of two add/drop units 3-16 and 3-17,
only one wavelength channel is extracted from the incoming
multiplexed main signal and fed to a fixed channel switch-type
optical switch 3-3. Provided in the optical switch 3-3 is a switch
SW for selecting signals input from the add/drop units 3-16 and
3-17. The add/drop units 3-16 and 3-17 can be constructed from
inexpensive dielectric filters and fiber Bragg reflectors. This
provides a cost-effective configuration for nodes that require only
a single channel add/drop capability.
(3) Node Example 3
[0109] FIGS. 14 and 15 respectively show main signal flows (1) and
(2) for an OADM local node in a logical star two-fiber UPSR optical
network.
[0110] The basic configuration in this example is the same as that
of Node Example 2 as shown in FIGS. 12 and 13, except for add/drop
capability to add one to three channels. In addition to the
add/drop units 3-16 and 3-17, this configuration also includes the
second add channel add/drop units 3-20 and 3-21 respectively for
the 0- and 1-paths and third add channel 0 and 1 add/drop units
3-22 and 3-23. As in Node Example 2, inexpensive dielectric filters
and fiber Bragg reflectors are used in these additional add/drop
units to provide a cost-effective configuration. If there is only a
short distance between the present node and the OADM units 3-14 and
3-15 (the nodes next to the present node in the 0-path and 1-path
fiber rings), and the loss between these nodes is small, no optical
amplifiers will be required at this point. In this case, the L-band
optical transmitters 3-18 and 3-19 of FIG. 15, which have no
amplifiers, can be substituted for the L-band optical
transmitter/amplifiers 3-10 and 3-11 of FIG. 14.
[0111] A second rack-mounting configuration is shown in FIG. 16.
The equipment required to be installed initially includes a
transponder rack 3-50, which includes an optical switch,
wavelength-division multiplexer/demultiplexer and a transponder
3-2-1. `0` and `1` add/drop units 3-16 and 3-17, `0` and `1` L-band
optical transmitter amplifiers 3-10 and 3-11, and an OSC (optical
supervisory channel) signal processor 3-60 are mounted on an
optical amplifier rack 3-52. Second add channel add/drop units 3-20
and 3-21 respectively for the 0- and 1-paths and third add channel
add/drop units 3-22 and 3-23 are added as required. In particular,
when a third add-channel capability is provided, an additional
optical switch/wavelength-division multiplexer/demultiplexer and
optical amplifier rack 3-52 is required. Since common equipment
types are used here and in the center node rack configuration shown
in FIG. 11, cost advantages are realized in terms of parts
availability, and reduced maintenance spares inventory.
(4) Node Example 4
[0112] FIGS. 17 and 18 respectively show main signal flows (1) and
(2) for an OADM local node in a logical star two-fiber UPSR optical
network.
[0113] The main signal flow configuration of this example is
essentially the same as that of Node Example 1. However, since it
is not necessary for this node to have add/drop capability for all
channels, as is required in the logical star-type center node of
Node Example 1, although additions are made in the future, in the
interim, for the unused channels, a direct optical fiber connection
is used from the wavelength-division demultiplexers 3-7 and 3-9 to
the automatic optical power level controllers 3-4 and 3-5.
[0114] If there is a long distance between the present node and the
OADM units 3-14 and 3-15 in respectively the 0-path and 1-path
fiber rings, and the transmission path loss between these nodes is
therefore large, optical amplifiers will be required at the inputs
of these nodes. In this case, the L-band optical
receiver/amplifiers 3-24 and 3-25 of FIG. 18 are alternatively
substituted for the 0-path and 1-path input bandpass filters 3-12
and 3-13 of FIG. 17.
(5) Node Example 5
[0115] FIGS. 19 and 20 respectively show main signal flows (1) and
(2) for an OADM center node in a two-fiber UPSR optical network.
This configuration differs from that of Node Example 1 in that a
dynamically switchable configuration is used in optical switch
3-26. In this configuration, rather than being dropped to the
transponder 3-2, the main signals input to this node from the ring
network are looped-back by the drop/through select optical switches
3-29 and 3-30 and add/through select optical switches 3-31 and 3-32
for transmission back into the ring. When this through-route is
selected, the additional optical power loss caused by these
switches is a problem. Referring FIG. 20, loss compensation
amplifiers (LCA) 3-27 and 3-28 are therefore respectively provided
in the 0- and 1-paths to compensate for this loss. This
configuration absorbs optical power level variations due to
switching.
[0116] Also, other functions are added by simply substituting
optical switches 3-26 for the optical switches 3-3 of Node Examples
1through 4. Since the different channels are all set to the same
power levels, different components and adjustment values are used
together. This enables `in-service expansion` in which optical
channels having the other functions are added with the remaining
channels maintained in a usable state.
[0117] FIG. 21 shows a rack mounting diagram for a central node
having an add/drop optical switch while FIG. 22 shows the rack
layout for a corresponding local node. In both of these racks, the
cables and connectors are provided to support the addition of
either standard optical switches 3-3 or add/drop/through optical
switches 3-26-9 through 3-26-16. This allows add/drop/through
switches to be used, as needed, in conjunction with the regular
switches, to provide functional expansion of service.
(6) Node Example 6
[0118] FIGS. 23 and 24 respectively show main signal flows (1) and
(2) for an OADM center node in a two-fiber UPSR optical network.
This configuration differs from that of Node Example 1 in that in
the optical switch 3-33 of this configuration, the main signals
input from the ring are divided and fed into two routes. One route
drops the signal to the transponder 3-2, and another route returns
it to the ring. The signal division is performed by drop/through
couplers 3-34 in the 0-path and 3-35 in the 1-path. As in Node
Example 5, additional optical power loss is contributed by the
additional switch in the through-route. Loss compensation
amplifiers (LCA) 3-27 and 3-28 are therefore respectively used in
the 0- and 1-paths to compensate for this loss.
[0119] Other functions are optionally added by simply substituting
optical switches 3-26 for the optical switches 3-3 of Node Examples
1 through 4. Since the different channels all are set to the same
power levels, different components and adjustment values are used
together. This enables `in-service expansion` in which optical
channels having the other functions are added with the remaining
channels maintained in a usable state.
[0120] As in Node Examples 4 and 5, FIG. 23 shows an example of the
main signal flow with no optical amplifier on the receive side
while FIG. 24 shows the main signal flow with an optical
receiver/amplifier 3-25 on the receive side.
3. Optical Fiber and Wavelength Band Range
(1) Dispersion-Shifted Optical Fiber
[0121] Described in this example is a C-band wavelength-division
multiplex system implementation using dispersion-shifted optical
fiber (DSF) transmission paths in which dispersion approaches zero
for wavelengths near 1550 nm (fiber in accordance with ITU-T G.653,
DSF). When a C-band or WDM carrier having equally-spaced
wavelengths is transmitted at normal light levels in a
dispersion-shifted fiber transmission path at a wavelength for
which dispersion is near zero, four-wave mixing occurs. A good
supply of inexpensive C-band optical components is already
available on the market. For example, the normal light level
conditions include -5 to 0 dBm for 1430 nm-1580 nm light in
accordance with STM16 of ITU-T G.957. The 1570 nm-1600 nm
wavelength band is referred to as L-band. When this band is used in
the DSF transmission paths equally-spaced wavelength placements are
possible for which dispersion becomes zero and disappears. This has
given rise to ideas for the use of L-band. Components for this
band, however, are in short supply and expensive.
[0122] Presented in this example is technology for using C-band
wave-division multiplexers in dispersion-shifted fiber transmission
paths. Four-wave mixing is a phenomenon that occurs when equally
spaced wavelength signal levels exceed -3.5 dBm per optical
channel. Because prior technology was directed toward high data
transition rates such as 2.5 Gbit/s and 10 Gbit/s at an optical
output of-3.5 dBm, it was not possible to obtain adequate
differences with respect to the minimum receive sensitivities based
on noise constraints. Thus transmission over practical distances
was not possible. For example, with a PIN photodiode used as a
detector, the minimum receive sensitivity was -18 dBm at 2.5 Gbit/s
and -14 dBm at 10 Gbit/s. There were problems in terms of optical
SNR (signal to noise ratio) constraints as well. For a path having
seven repeater optical amplifiers with NF (noise figure) of 7 dB,
the minimum receive sensitivity was -24 dBm at 2.5 Gbit/s and -18
dBm at 10 Gbit/s. Also, the best value that could be achieved for
maximum path loss in the fiber transmission path with compensation
is 14 dB at 2.5 Gbit/s, 8 dB at 10 Gbit/s, 10-14 dB at 2.5 Gbit/s
and 4-8 dB at 10 Gbit/s with tolerances applied. This made it
difficult to find practical applications for this technology.
[0123] The Gigabit Ethernet data transmission rate is 1.25 Gbit/s.
At this data rate, the SNR is 3 dB better than that at 2.5 Gbit/s,
which enables a compensated optical fiber transmission path loss of
12 dB. Thus with margin, this makes a C-band 20-40 km optical fiber
transmission path possible. The same transmission is difficult to
do at 10 Gbit/s) possible.
[0124] The present example is a method for transmitting a 1.25
Gbit/s data rate optical signal in a DSF optical fiber transmission
network with seven repeater nodes separated by 20-40 km spans, in
which the four-wave mixing that characteristically occurs in DSF at
C-band is avoided by reducing optical power levels. With the span
loss is on the order of 12 dB, one optical amplifier repeater stage
at each node either as a preamp or a postamp is sufficient. The
function of the repeater optical amplifier at each node is to
amplify optical channel signals passing through the node, but
optical channel signals dropped at that node must also be accounted
for. Because reducing optical signal levels could result in
insufficient input to optical receivers, in the present example,
optical preamplifiers are used to ensure adequate levels at the
inputs to optical receivers. If the required output level of the
node ranged from -14 to -3.5 dBm, the span loss would be 12 dB, and
the transmission path penalty would be 1 dB. The minimum optical
receive level at the node input would be -27 dBm. This level is
amplified by the optical preamplifier to -20 dBm at the receiver
input so as to provide a margin of 10 dB with respect to an optical
receiver unit receive level of -30 dBm.
(2) Normal Dispersion Optical Fiber
[0125] In this example, C-band wavelength-division multiplex system
implementation is described using a normal-dispersion optical fiber
transmission path in which dispersion approaches zero for
wavelengths near 1310 nm in accordance with ITU-T G.652, SMF.
Unlike in a dispersion-shifted transmission path, because there is
dispersion on the order of 17 ps/nm/km at wavelengths near 1550 nm,
a C-band multiplex carrier with equally spaced wavelength channels
is transmitted with no concern for four-wave mixing regardless of
the optical signal level. An abundant supply of inexpensive C-band
optical components is already available on the market. For the
system of the present example, a span length of 20-40 km was
assumed with span loss of approximately 12 dB. Because an adequate
SNR margin is provided by using high optical power levels, up to 16
nodes are included, and with signal conversion the configuration is
expanded up to 15 optical repeater nodes. With a span loss range of
approximately 12 dB, a single optical pre or a post amplifier at
each repeater node is enough. The function of the repeater optical
amplifier at each node is to amplify optical channels passing
through the node. Since the number of necessary amplifiers
corresponds to the number of nodes, this is an area in which it
would be desirable to control costs. In direct contrast to the DSF
above discussed example, in which optical receiver constraints were
a concern, at high optical levels, it is the constraints on output
level at the transmit end. In general, optical transmitters become
substantially more difficult to manufacture and thus are
substantially more expensive when they are made for optical output
levels of 0 dBm or more. Between the optical transmitter and the
node output, the signal passes through a multiplexer, ALC, and
demultiplexer units, and the signal encounters approximately 9 dB
of loss. For this reason, postamplifiers rather than preamplifiers
are used. If the overall optical level is high, the levels applied
to the receivers of the nodes will also be high, and preamplifiers
will not be required.
4. An Optical Fiber and a Wave Length Range
(1) A Dispersion Shifted Optical Fiber
[0126] Referring to FIG. 25, an embodiment of a C-band multiplex
wave length transmitting system to a dispersion shifted optical
fiber (stipulated in the ITU-T G.653, DSF) transmitting path with a
wave length in the vicinity of 1550 nm shows a chromatic
dispersion. FIG. 25 shows a four wave mixing phenomenon when
optical signals of the C-band which are at equal spacings are
transmitted to a DSF transmitting path. In the dispersion shifted
optical fiber (DSF) transmitting path 3-74, when optical signals of
a wave length of the C-band which show a substantially zero
chromatic dispersion are arranged at equal wave length spacings as
follows: when .lambda.1=optical signal 3-71, .lambda.2=optical
signal 3-72, and .lambda.3=optical signal 3-73, are input, the four
wave mixing phenomenon is caused, resulting in an output of wave
length .lambda.1-.DELTA..lambda.=optical signal element 3-84 and
wave length .lambda.2+.DELTA..lambda.=optical signal element 3-85
as a four wave mixing element between wave length .lambda.1=optical
signal 3-81 and wave length .lambda.2=optical signal 3-82 in
addition to wave length .lambda.1=optical signal 3-81, wave length
.lambda.2=optical signal 3-82 and wave length .lambda.3 optical
signal 3-83. The above mentioned .DELTA..lambda. is the difference
between the wave length .lambda.1 and the wave length .lambda.2.
The wave length .lambda.3=optical signal 3-83 which is arranged at
equal spacings interferes with the wave length
.lambda.2+.DELTA..lambda.=four wave mixing optical signal element
3-85, resulting in deterioration of the signal to noise ratio of
the wave length .lambda.3 optical signal. The optical component of
the C-band has been already commercially available, has a low cost,
and is supplied sufficiently. So far, there has been an idea to use
a wave length range from 1570 to 1600 nm or the L-band because the
wave length of the L-band does not show any chromatic dispersion
and is arranged at equal spacing when it is applied to the DSF
transmitting path. However, when the optical component of the
L-band is not supplied sufficiently, the length of a fiber used to
amplify is large in principle, and the use of the L-band needs a
high cost to introduce a mechanism which avoids a reduction of the
excitation efficiency or a temperature dependency.
[0127] In the present embodiment, the method to apply the multiplex
wave length of the C-band to the DSF optical fiber transmitting
path is described. The four wave mixing is a phenomenon that, when
the wave length at equal spacings is -3.5 dBm or more than -3.5 dBm
per optical channel, an adjacent wave length element generated by
the four wave mixing becomes larger, resulting in a deterioration
of the signal. The conventional optical fiber transmitting path has
used a high rate such as 2.5 Gbit/s and 10 Gbit/s. Therefore, an
optical output of -3.5 dBm does not have a difference from the
minimum reception sensitivity due to the noise restriction,
resulting in an impractical transmitting distance. For example,
when pin-PD is used, the optical output is -18 dBm at 2.5 Gbit/s
and -14 dBm at 10 Gbit/s. Considering the restriction due to the
signal to noise ratio (SNR or S/N), when seven relay optical
amplifiers with noise figure (NF) of 7 dB are used, the minimum
reception sensitivity is -24 dBm at 2.5 Gbit/s and -18 dBm at 10
Gbit/s. In the optical fiber transmission loss, the maximum
compensation is 14 dB at 2.5 Gbit/s and 8 dB at 10 Gbit/s, and in
consideration of a dispersion of 4 dB, the compensation is 10 dB at
2.5 Gbit/s and 4 dB at 10 Gbit/s at most, resulting in an
impractical optical fiber transmitting path.
[0128] The transmitting rate of Gigabit Ethernet is 1.25 Gbit/s.
Since the S/N in the rate of 1.25 Gbit/s is 3 dB better than that
in the rate of 2.5 Gbit/s, the compensation of the optical fiber
transmission loss is 12 dB. Therefore, the C-band optical fiber
transmission of 20.about.40 km is possible with a small margin by
using Gigabit Ethernet, but not by an optical fiber transmission
system with the rate of 10 Gbit/s.
[0129] The exemplary embodiment according to the current invention
describes, the transmitting method in the DSF optical fiber
transmitting network transmitting an optical signal at the rate of
1.25 Gbit/s relayed via seven nodes at span intervals of
20.about.40 km, in which the optical level is reduced to avoid the
four wave mixing specific to DSF caused in the C-band. When the
span loss is around 12 dB, one relay optical amplifier is enough
for one node. Either a front optical amplifier or a rear optical
amplifier is used. The relay optical amplifier functions as an
optical channel which passes through the node. Considering that the
optical level of the optical channel drops within the node, the
present embodiment uses a front optical amplifier to ensure the
level of the optical input to an optical receiver because the
optical input to the optical receiver is not enough when the
optical level is reduced. In other words, the present embodiment
has a margin of 10 dB for the reception level of 30 dBm per unit in
the optical receiver, when the optical input to the optical
receiver is increased to more than -20 dBm by amplifying the
minimum level of the optical reception of -27 dBm using the front
optical amplifier on condition that the maximum optical output from
the node is -3.5 dBm, the minimum of that is -14 dBm, the span loss
is 12 dB and the transmitting path penalty is 1 dB.
[0130] The present invention provides an optical network which has
a common component without depending on a type of a signal
connected to a client, a type of a transmitting path and a wave
range used. When a dispersion shifted optical fiber is used in an
optical fiber transmitting path, a wave length which shows no
chromatic dispersion is in the vicinity of 1552 nm. Therefore, when
wave lengths within a wave length range of the C-band at equal
spacings are used for the conventional optical fiber transmitting
path with a distance of more than 40 km stipulated in the ITU-T,
the cost of an optical component is inexpensive. However, the
interference mode between two wave lengths called a four wave
mixing places the waves with the wave length above other signals or
other waves at equal spacings, resulting in a phenomenon where an
optical transmission characteristic is deteriorated. In a multiplex
wave length transmission using a dispersion shifted optical fiber,
one object is how to use inexpensive optical components for the
C-band. The method named an unequal spacing has been described
previously to remove the influence by the above mentioned
phenomenon. However, the reduction of the high cost of an optical
component used in the method was not achieved because the design of
the optical component was complex. It is important to solve the
above mentioned object.
[0131] The concern here, however, is the level input to the
postamplifier. When the input level applied to an optical amplifier
increases, greater optical excitation power is required to obtain
the same gain. However, because the function of an optical
amplifier in a transmission system design is to compensate for
loss, excessively high optical level inputs are wasteful and tend
to increase cost. In a 16-channel optical transmission path with
20-40 km spans, the span loss is approximately 12 dB. In such a
system, an optical postamplifier with one excitation light source
[(pump)] and a gain of approximately 20 dB with an optical input
level of -20 dBm to -17 dBm will cause a high total system SNR in a
configuration that is expandable up to 20 nodes. At higher optical
levels beyond the above described levels,
two-excitation-light-source optical postamplifiers are required.
The network alternatively need to limit to less than 16 nodes if
one excitation light source is used.
[0132] The above provided present example is a highly
cost-effective system for transmitting 16-channel C-band
wavelength-division multiplex signals in an SMF fiber transmission
path with 12 dB span loss. The system uses comparatively low-output
light sources and one-excitation-light-source optical amplifiers
that are used as optical postamplifiers in each node of the 16-node
system with amplifier per-channel input levels of -20 to -17 dBm.
Amplifier gain is set to approximately 20 dB.
(2) Mixed Optical Fiber
[0133] A lower-cost C-band wavelength-division multiplex optical
network is configured using a combination of G.652 and
G.653-compliant fibers by using the optical levels specified in the
above paragraph (1) for dispersion-shifted optical fiber.
[0134] According to the present invention as described above,
medium-scale optical networks for IP communications are configured
at a low cost: optical networks provide a steady cash flow due to a
lower initial capital investment but are easily expandable through
addition of facilities. Also, according to the present invention,
optical networks use common components regardless of the kinds of
transmission paths or wavelength bands. In addition, according to
the present invention, in addition to making effective use of open
transmission paths, a highly reliable two-fiber network is
supported in the optical layer.
* * * * *