Over the past two decades, technological and economic forces have driven the telecommunications world. Deregulation and privatization have combined with improvements in optical-fiber technologies to make data and voice traffic grow at exponential rates. Figure 1 shows how network capacity has changed from the early 1980s to the year 2000.
In the 1990s, the high volume of both data and voice traffic brought about the use of wavelength division multiplexing (WDM) in which several frequency (or wavelength) signals are simultaneously transmitted on a single fiber. In the late 90s, modular dense WDM (DWDM) became a reality when multiple groups of services and multiple wavelengths per group were introduced.
Enabling Technologies
With technologies such as optical amplifiers, DWDM boosts network transmissions and provides more bandwidth. When erbium-doped fiber amplifiers (EDFAs) were introduced, they helped to reduce space and power requirements and brought down the cost of construction. With the development of new EDFAs operating in the L-band (1,570 nm to 1,610 nm), Raman fiber amplifiers (RFAs), and improvements in solid-state amplifiers, new bandwidth ranges will become available.
Another technology that increases the use of DWDM is the optical add/drop multiplexer (OADM) which is especially suited to adding and dropping traffic at smaller sites. This opens up DWDM transmission in the metropolitan areas where connectivity is a key issue compared to long-haul transmissions where cost per managed bit-kilometer is crucial.
Today’s OADM technology supports four to eight add/drop channels. The next generation should offer a flexible combination of channels with the use of tunable lasers.
A third technology that enables DWDM is switches and routers such as optical cross-connects (OXC) used for protection restoration and routing. New material technologies will make these routers available in massive M × N switching capabilities.
Network of the Future
With all these technologies, each DWDM network application then will be connected to a unique wavelength. Optical fibers will have to support a large number of wavelengths without being impaired by limiting phenomena such as polarization mode dispersion (PMD) and nonlinear effects.
Consequently, future networks will be characterized by data centricity and service granularity. Data centricity relates to building networks to address data- transmission requirements, given that data traffic will continue to grow at a tremendous rate. Granularity relates to the flexibility of a system, where individual items in a flow of information are broken down into relatively small packets. The finer the granularity (the smaller the packet), the greater the potential for speed and flexibility.
The services offered will be characterized by DS1 (digital service) and DS3 voice and private lines, OC-3 (optical carrier) and OC-12 private lines, OC-12 and OC-48 asynchronous transfer mode (ATM) trunk lines, OC-3 to OC-48 Internet trunks, and video and Gigabit local area network (LAN) transmissions, each on its own wavelength. And OC-192 and OC-768 eventually will replace OC-48. Table 1 illustrates a time line of the technology for DWDM network transmissions, assuming optics will stay ahead of demand for capacity.
DWDM, however, will not provide all the answers to the continuous and insatiable demand for bandwidth. In theory, the total bandwidth available on a single-mode optical fiber is about 400 nm (50 THz), assuming loss of <0.5 dB from 1,225 nm to 1,675 nm. But in practice, we can realistically expect a 400-km total capacity of 2.5 Tb/s in about five years and 20 Tb/s in about 10 years.
A 1-Tb/s rate can be achieved by using 400 wavelengths at 2.5 Gb/s with a minimum optical signal-to-noise ratio (OSNR) of about 20 dB, or 25 wavelengths at 40 Gb/s using a minimum OSNR of about 32 dB. But as a general rule, the higher the bit rate per channel, the greater the problems for system designers. So 2.5 Tb/s or 20 Tb/s total bandwidth capacity will not be easy to achieve.
DWDM still will provide considerable space saving in repeater hubs but at the expense of an increased density of transmitters in the central office. It will continue to be the accepted transport technology for interexchange carrier networks. Worldwide DWDM networks will continue to be constructed with combinations of land and undersea cable deployments.
Very soon, DWDM will use fixed OADM including optical protection (wavelength and span) for partially flexible bus and ring transmissions. In the short term, flexible OADMs will be used. In the medium term, OXC will be used in meshed networks. To support such a development, DWDM networks will face several key challenges including tighter controlled laser tolerance, tighter optical filter tolerance, nonlinear effects, complex power budget engineering, and optical noise accumulation.
Migration toward an Internet Protocol (IP) optical networking backbone with layered bandwidth management will be used to achieve lower cost per managed bit, fewer layers, optical layer protection and restoration, optical reconfiguration, and maximum use of data bandwidth. Several options will be available to the networks, such as new fibers, OC-192 and higher, switching, ring and mesh structures, and optical add/drops and cross-connects.
Key technologies will be wavelength switching, interchange, and conversion; low-cost dispersion compensation; and all-optical regeneration. These technologies indicate that the market for DWDM equipment and components certainly will continue to grow even if the involved technologies become increasingly complex.
Several developments should be expected in DWDM networks by 2010, provided that the bandwidth doubles every six months. Cables and fibers will be denser, with up to 864 fibers per cable—except in countries such as Japan where the infrastructures favor much larger fiber capacity per cable—and up to 128 wavelengths per fiber.
Standards also are in progress for the optical layer of DWDM networks. Table 2 summarizes International Telecommunications Union (ITU) recommendations on DWDM.
The new optical network will assume retimed, reshaped, and reamplified (3R) regeneration at each subnetwork. As transmitters with tighter tolerances will be required, cost, floor space, power, sparing, network management, and training will become critical issues. The key solution will be the use of tunable lasers.
The next generation of transponders will have full optical service channel functionality, full synchronous optical network (SONET) overhead processing and signal conversion without multiplexing, enhanced modulation techniques and formats, out-of-band framing, forward error correction (FEC), data channel communications, automatic protection switching, bit-error performance monitoring, and signal type/trace information.
The next generation of SONET will offer significantly lower cost and improved density as well as optimized tributary interfaces for voice and data, tunable ITU-compliant transmitters, in-band FEC, cost-effective AC and DC power management, remote testing, and simplified user access. Floor space, size, and power dissipation will be issues. We also will see the tendency to have IP over SONET because of its simplicity, bandwidth efficiency, scalability, network optimization for IP traffic, fault tolerance (with SONET routers), and a single structure to manage.
To facilitate the long-term development of an all-optical network, we will see the use of several technologies including higher channel-count OADMs, optical 3R criteria, ultra-fast and dense optical switches, and new types of performance monitoring and management. RFAs will be used with EDFAs to extend the network span, as shown in Figure 2.
Characterized by return-to-zero (RZ) modulation with narrow pulses traveling in groups, solitons will
provide longer propagation distance, lower sensitivity to nonlinear effects, and tolerance to PMD and pulse broadening, but will require accurate dispersion management.
Conclusion
With the business development of new services such as the Internet and e-commerce, telecommunications networks face new challenges. DWDM systems have established the path for a successful response to the continuous demand for more services, capacity, and bandwidth. There certainly is a tendency to bring IP over SONET, which also is defining the path toward an all-optical network.
It still is not clear if this all-optical network will be established in the foreseeable future because many experts still believe that the transmitted bit of information will have to be electronically monitored, tested, and managed. However, the present growth in DWDM component, subsystem, and system technical development is a clear indication that the race for an all-optical network is wide open.
About the Author
Dr. Andre Girard, the senior member of the technical staff at EXFO, has more than 28 years of experience in fiber optics, lasers, and systems/project engineering. After joining EXFO in 1994, he launched the company’s Scientific Division and has participated in international standards organizations. Dr. Girard earned his Ph.D. in physics from the National Institute for Scientific Research in Montreal in 1979. EXFO,
465 Godin Ave., Vanier, Quebec Canada G1M 3G7, (418) 683-0211, e-mail: [email protected].
1998
2000
2002
Fiber
· NZDSF
· Dispersion compensation
· Dispersion managed fiber
· Better compensation
· Lower loss
· Multicore fiber
· Ultra-low loss
· Solitons
Transmission
· 0.5 Tb/s
· 2.5 Gb/s per l
· DWDM 32 l
· 1.5 Tb/s
· 10 Gb/s per l
· DWDM 128 l
· 2 to 5 Tb/s
· 40 Gb/s per l
· Soliton
· UDWDM 200+ l ‘s
Bandwidth Management
· Electrical
· Async, SONET
· Interfaces <10G
· STS granularity
· Electrical, optical
· Multiprotocol, ITU l grid
· Interfaces up to 10G
· STS granularity
· DCS
· Optical
· Multiprotocol, ITU l grid
· Interfaces up to 40G
· STS-3c granularity
· Fast switching
Table 1.
Table 2
Rec.
ex
Q
SG
Status
Title
G.661
17
15
A
Definition and Test Methods for the Relevant Generic Parameters of Optical Amplifier Devices and Subsystems (10/98)
G.662
17
15
A
Generic Characteristics of Optical Amplifier Devices and Subsystems (10/98)
G.663
17
15
A
Application Related Aspects of Optical Fibre Amplifier Devices and Subsystems (10/96)
G.664
G.saf
16
15
D
General Automatic Power Shut-Down Procedures for Optical Transport Systems (10/98)
G.665
G.onc
17
15
W
Optical Network Components and Subsystems
G.671
17
15
A
Transmission Characteristics of Passive Optical Components (11/96)
G.681
9
15
A
Functional Characteristics of Interoffice and Long-haul Line Systems Using Optical Amplifiers Including Optical Multiplexing (10/96)
G.691
16
15
D
Optical Interfaces for Single-Channel SDH Systems with Optical Amplifiers (6/98)
G.692
16
15
A
Optical Interfaces for Multi-Channel Systems With Optical Amplifiers (10/98)
G.709
G.ons
11
15
–
Network Node Interface for the Optical Transport Network
G.798
G.oef
9
15
–
Characteristics of Optical Transport Network Hierarchy Equipment Functional Blocks
G.871
20
15
–
Framework for Optical Transport Network Recommendations
G.872
G.otn
19
13
A
Architecture of Optical Transport Networks (2/99)
G.873
G.onr
19
13
–
Optical Transport Network Requirements
G.874
G.onm
13
15
–
Management Aspects of the Optical Transport Network Element
G.875
G.oni
14
15
–
Optical Transport Network (OTN) Management Information Model for the Network Element View
G.959.1
G.onp
16
15
–
Optical Networking Physical Layer Interfaces
A = Approved; D = Determined; – = Draft Text only; ex = Interim Designation; Q = Question; SG = Study Group
Copyright 1999 Nelson Publishing Inc.
October 1999
|
|