Carter Houghton
Until recently, dense wavelength-division multiplexing (DWDM) and erbium-doped fiber amplifiers (EDFAs)-essential elements in today`s long-haul communication networks-were not widely accepted as cost-effective solutions for metropolitan networks. Installing new fiber or upgrading SONET equipment seemed to be the most cost-effective ways of expanding network capacity, and metro networks were not expected to be large enough to require amplification.
This scenario is changing rapidly as system houses look to build transparent and scalable networks that can handle different protocols, bit rates, and traffic patterns. Metro DWDM, once considered too costly, is making an economic case for itself in some metropolitan areas. As a result, demand for metro EDFAs is beginning to surge for two important reasons. First, it`s clear that some ring distances are long enough to require amplification. Second and even more important, EDFAs are being called upon to compensate for component loss that arises from using various DWDM technologies.
Although the metro market will require fewer amplifiers than the long-haul market, at least in the near term, the variety of amplifiers required by the market could be much greater because of the large number of potential applications. Metro amplifiers will have to strike a balance between standardization, which reduces cost and development time, and customization, which allows amplifiers to meet the demands of various applications and architectures. The next 12 months will be critical to the evolution of metro networks and EDFAs as system houses move beyond field trials into actual deployment.
Defining a metro network
Metro networks manage traffic within a local, metropolitan, or regional area, as well as route traffic directly to and from the backbone. The networks are generally bidirectional ring, mesh, or point-to-point architectures. The bidirectional ring will likely emerge as the dominant metro architecture in the coming years because it offers more flexibility and protection compared with other architectures. Today, three primary types of metro DWDM networks exist: regional/interoffice (IOF), access, and enterprise (see Fig. 1).
Regional/IOF networks act as feeder networks and typically link the interoffice facilities of local exchange carriers and the points-of-presence of inter-exchange carriers. Regional/IOF networks have six to 12 nodes on average and can range from 100 to 300 km in circumference. They require amplification both to boost signals and to compensate for component loss. Regional/ IOF networks are currently driving the need for amplification in the metropolitan market.
Access networks distribute traffic from a regional/IOF network to end-users such as Internet service providers and competitive local-exchange carriers. A typical access network has five to eight nodes and can range from 20 to 100 km in circumference. Amplification in access networks is currently limited, but that trend could change down the road as ring circumferences, node counts, and bit rates increase.
Enterprise networks link different sites within organizations such as large corporations, financial institutions, and universities. They are typically 5 to 20 km in circumference and generally require no amplification.
Connections between regional/IOF, access, and enterprise networks are not optical. For example, a signal traveling from one access ring to another goes through an optical-to-electrical-to-optical (O-E-O) conversion at an electrical crossconnect as it travels onto and off of a regional/IOF ring. Emerging technologies such as optical crossconnects are expected to decrease the number of O-E-O conversions necessary in networks. This trend will increase the need for amplification in metro networks because a signal will travel through multiple rings without going through an O-E-O conversion and therefore will experience greater attenuation.
EDFA applications in metro networks
Today`s metro DWDM networks, which are in their infancy, use several types of EDFAs (See Fig. 2).
DWDM ring/line amplifier: In some metro rings and point-to-point applications, EDFAs are used to boost signals because of span loss. DWDM ring/line amplifiers normally handle from 16 to 32 channels. They are generally used in regional/IOF networks and occasionally in large-access networks.
WADM pre/post amplifier: Wavelength add/drop multiplexers (WADMs) are used in metro networks to add and drop channels at a node. To compensate for component loss associated with WADMs, a preamplifier can be used before the node and/or a postamplifier can be used after it. In some cases, one amplifier (a pre- and post-amplifier) combined with mid-stage access can be used. Typically, WADM pre/post amplifiers handle from 16 to 32 channels and include embedded control electronics to handle power fluctuations that arise from the adding and dropping of wavelengths. They are generally used in regional/IOF and access networks.
WADM band add/drop amplifier: These amplifiers can be used in regional/IOF and access networks where the majority of channels pass through a node unamplified. At each WADM, however, a specific set of channels referred to as a "band" is added or dropped onto the fiber. The WADM band add/drop amplifier boosts only the band that is added or dropped. For dropped channels, it functions as a band preamplifier prior to demultiplexing into the receivers. For added channels, it functions as a band postamplifier following the multiplexing stage after the transmitters. Typically, WADM band add/drop amplifiers handle up to four channels.
Single-channel amplifier: In some low-channel-count applications, amplifying individual wavelengths at the transmitter or receiver is preferable because of cost. Single-channel amplifiers, new to metro networks, are mainly used in access networks.
Tomorrow`s metro amplifiers
Metro network operators will face a number of trade-offs when considering performance, cost, and availability of metro amplifiers. This balancing act will likely lead to a variety of amplifiers developed specifically for metro applications.
While the amplifiers described above will continue to make up a large part of the metro EDFA market, new types of amplifiers are on the horizon. For example, optical networking devices such as crossconnects and switches will require amplifiers. Some of these amplifiers will be integrated into the design of the devices. For the moment, optical networking devices are largely in development, and it is unclear whether they will incorporate single low- or high-channel amplification and what performance attributes they will have.
EDFAs are clearly the most trusted technology for amplification in metro networks. For high-performance applications, metro EDFAs such as DWDM ring/line amplifiers and WADM pre/post amplifiers are the technology of choice. Some metro applications may lend themselves to single-channel metro EDFAs or other emerging technologies because of higher cost sensitivity and lower performance requirements (see "The reality behind low-cost amplifiers," p. 28). Semiconductor optical amplifiers (SOAs) and planar waveguide amplifiers may prove to be alternative technologies for such applications.
Contrary to earlier perceptions, the metro DWDM market does require amplification. While demand today for amplifiers numbers in the thousands, that figure could increase sharply as DWDM becomes more widespread in the market and optical networking technologies penetrate metro architectures.
Although metro amplifiers are often referred to as scaled-down versions of their long-haul cousins, they are often designed with unique features such as embedded control electronics. In fact, the metro amplifier market is proving to be more diverse and competitive than originally thought. The challenge to amplifier manufacturers will be not only to develop and manufacture many different types of metro amplifiers but to do so quickly and cheaply.
Carter Houghton is a product line manager for optical amplifiers at Corning Incorporated, Photonic Technologies Division, One Riverfront Plaza, Corning, NY 14831. He can be reached at 607-974-4229; fax: 607-974-7827; or by e-mail: [email protected].FIGURE 2. Potential applications for EDFAs, each with different performance requirements include add/drop multiplexing, line amplification, and single-channel amplification.
The reality behind low-cost amplifiers
The industry has witnessed a lot of activity lately around low-cost amplification in the metro market. At this year`s Optical Fiber Communication conference (Baltimore, MD; March 7-10), a session on sub-$1000 amplifiers was packed with employees from components suppliers, system houses, and service providers alike. Several companies have introduced amplifier solutions specifically targeted to the metro market.
The prospect of low-cost amplification is of particular interest to the metro market because it is extremely price-sensitive. Amplifiers that are typically deployed in long-haul and undersea networks are too costly for metro networks, which operate in a more competitive environment and which allocate costs over much shorter distances.
The phrase "low-cost amplification" is widely used in the metro market today, perhaps too widely. In fact, it is somewhat of a misnomer. Everyone wants low-cost amplification regardless of the market segment, amplifier type, or application. But are companies willing to give up performance for lower prices? Amplifier makers, in order to meet performance levels specified by their customers, must incorporate certain components and technologies into their designs-and that comes at a price. Standard DWDM amplifiers used in current metro networks-DWDM ring/line amplifiers and WADM pre/post amplifiers-are examples of higher-performance metro amplifiers for which network operators are less likely to sacrifice performance for price.
Amplifier manufacturers recognize that many different types of amplifiers are and will be needed in the metro market. Metro DWDM amplifiers will be used for the applications mentioned above. Additionally, new types of metro amplifiers will be developed. For example, single-channel amplifiers are gaining acceptance in the market because of the perceived cost savings they might offer. When the industry refers to "low-cost amplification" in metro, it is likely referring to single-channel amplifiers.
It is much less costly to make a single-channel amplifier than it is to make a metro DWDM amplifier. Designs incorporate lower-power pump lasers, which cost less than higher-power lasers typically used in metro DWDM amplifiers. In addition, single-channel amplifier designs do not require as many expensive optical components, in particular gain-flattening filters and, in some instances, taps and photodiodes.
However, single-channel amplifiers, which cost on average $3000, must become much less expensive if they are to prove themselves in metro networks, especially because a simple metro DWDM amplifier costs around $10,000. Looking at cost on a per-wavelength basis, a single-channel amplifier today would only justify itself in very low channel count (less than four channels) applications.
Three cost-cutting factors
What will it take to drive down costs of single-channel amplifiers? These amplifiers need to maintain a certain level of performance while increasing the cost gap with metro DWDM amplifiers. Manufacturing in large volume can bring down costs rapidly. However, large-volume manufacturing generally requires some level of standardization, and it is still unclear as to whether systems houses are willing to embrace standard product solutions.
A second cost-cutting factor is new component technologies. Uncooled pump lasers are an emerging technology receiving interest in the market today. However, these devices alone will not drive costs to where they need to be. Other optical components, which can make up a large percentage of the total cost of a single-channel amplifier, need to become less expensive as well.
The third cost cutter is a technology breakthrough. Semiconductor optical amplifiers (SOAs) and planar waveguide amplifiers could one day offer lower costs than traditional EDFAs as well as other benefits, including smaller packaging. But neither technology is currently deployed and both need to improve their performance to become a mainstay in networks. For example, SOAs will not become widely deployed until they reduce noise figures from current levels (9 to 10 dB) down to 6 dB or less.
So what does all this mean? It appears that single-channel amplifiers, both EDFAs and emerging technologies, will likely find a home somewhere in metro networks. They cost about 30% of metro DWDM amplifiers. To become widespread in the market, single-channel amplifiers will need to cost less than 10% of metro DWDM amplifiers. The metro market is much farther away from sub-$1000 amplifiers than is realized, and it appears that metro DWDM amplifiers will dominate the landscape in the near term.