wdm technology enriches sales, networks, and investments
The fast adoption of wavelength-division multiplexing technology by both telephone companies and equipment vendors confirms that multichannel, all-optical fiber networks are viable and lucrative market endeavors
john p. ryan
ryan, hankin, kent inc.
The adoption of erbium-doped fiber amplifiers (edfas) in the early 1990s represented the first step toward all-optical fiber networks and led the way for acceptance of wavelength-division multiplexing (wdm) technology. The key economic factor that makes wdm networks the favored technology for long-distance carriers is that all wavelengths in the 1550-nm window can be boosted by a single amplifier. Of course, an upgrade from OC-48 (2.5 Gbits/sec) to OC-192 (10 Gbits/sec) can also benefit from the use of optical amplifiers. However, if this upgrade path is chosen instead of implementing wdm technology, it is restricted to one bitstream.
Beneficial economics have therefore propelled dense wavelength-division multiplexing (dwdm) technology from the research laboratory to commercial long-distance carrier applications in record time. For example, at&t Corp. purchased 4-wavelength systems from Lucent Technologies in late 1995. During 1996, Sprint Inc. awarded wdm contracts to Pirelli Telecom Systems and Ciena Corp. In addition, WorldCom Inc. bought Ciena`s equipment late in 1996. And early this year, mci Communications Corp. started using Pirelli equipment in conjunction with Hitachi OC-192 systems. Furthermore, dwdm long-distance networks are spreading internationally, with trials starting or under way, and contracts expected, in Britain, Spain, and Germany.
ibm`s 10-wavelength system spurred large sales during 1996. Rather than being used for long-span interexchange carrier routes, the ibm system targets short-haul links across town, campus links of a single organization, and leased dark-fiber links.
The aggregate dwdm business of Lucent, Ciena, Pirelli, and ibm, as well as their different approaches to using wdm technology, totaled more than $350 million in 1996, a sevenfold increase from 1995. By the year 2000, the dwdm market is forecast to exceed $1 billion for interexchange carriers.
In 1997, wdm systems are being installed in field trials for several regional carriers and competitive access providers. These telecommunications service providers are all exploring the potential business resulting from the use of dwdm in short-haul networks.
Investor interest
Important business developments have also focused on financial investments in wdm technology and products. In fact, they have established a new market parameter: Optical-network technologies hold the potential of skyrocketing market sales. Here are some corroborating developments:
Ciena`s initial stock offering last February was the largest in the history of fiber optics and telecommunications. Its market capitalization of $3.7 billion was the largest in the history of startup companies. This capitalization is attributed to Ciena`s executives; their timing of introducing dwdm technology helped accelerate the market for optical networks (see "WDM network operation calls for close control of parameters," page 27).
Newbridge Networks, best known for its Asynchronous Transfer Mode and frame-relay systems, made investments in two optical-network subsystem companies. One investment involved Bookham Technology, a British firm whose in-silicon wave guide technology has the potential of integrating multiple elements for optical-network systems. The second investment was made in a Canadian startup company called Cambrian, which focuses on wdm components.
Tellabs has taken over the public network rights to the wdm technology developed at ibm`s Thomas J. Watson Center under the direction of Paul Green (see Lightwave, April 1997, page 1). Consequently, the original developer of the Rainbow 2 all-optical network (see Lightwave, January 1995, page 1) and the MuxMaster 20-channel fiber-optic multiplexer (see Lightwave, March 1995, page 1), becomes two entities: ibm continues to develop the technology for its own purposes, and Tellabs explores optical-networking technology to complement its own communications systems.
Corning Inc. has acquired Massachusetts-based Optical Corporation of America, a company that manufacturers wdm components and other products. These components are expected to support a growing optical-network product suite at Corning, which already embraces optical-amplifier elements.
All these wdm business transactions, which range in market value from $6 million to $4 billion, have occurred in just seven weeks during the first quarter of this year. Moreover, they all reaffirm that strong investor interest has emerged this year for wdm and optical-network technologies
High-bandwidth demand
Presently, optical networks represent a growing market for the same reason that some major long-distance carriers are no longer able to sell DS-3 44.736-Mbit/sec lines--users` demands for bandwidth capacity on the nation`s long-distance backbone networks are at record levels and rising briskly. Unfortunately, long-distance carriers are running out of network capacity for the following reasons:
Frame-relay service revenues are increasing at a rate of 30% per year. Indeed, the communications traffic underpinning this rate is rising even faster.
Internet traffic is continuing to rise at an astonishing rate--between 6% and 10% per month. This expansion means that the capacity of the entire Internet needs to be duplicated within one year.
Voice traffic also continues to grow, so that increased competition is expected to decrease prices, which, in turn, will have a marked impact on call volumes.
This surge in bandwidth demand translates into a flourishing communications business market. On the other hand, it means serious network capacity problems for major interexchange carriers: They do not have the capacity needed to handle this unending new wave of traffic (see table). Note that Sprint and mci are running close to 85% active fibers as an average across their entire networks. These numbers mean that the major routes for both carriers are nearly 100% active, with no sign (or expectation) of abatement in user demand for capacity.
WorldCom has reported fewer active fibers; its aggressive fiber build-out program means that the final 1996 data may be lower than listed. This carrier is also experiencing severe capacity shortages on the East Coast. Stentor, which runs the major Canadian national telecommunications network, also has large spans that are fully activated, with no extra capacity available.
Fortunately, some bandwidth solutions are available for the major interexchange carriers:
Plow in new fiber-optic cables on major routes. However, this approach is time-consuming and costly, and the major interexchange carriers are showing no signs of wanting to undertake this venture.
Lease capacity from other network providers, especially secondary carriers. Major interexchange carriers are likely to take this route for the short term.
Upgrade networks via time-division multiplexing (tdm) or wdm. The major interexchange carriers are focusing on wdm technology because it can provide capacity growth beyond the current tdm upgrade from OC-48 to OC-192. With wdm, multiple tdm wavelengths or channels are combined on a single fiber to increase the effective capacity more than the fourfold increase possible using tdm only.
Photonics approach
So far, these bandwidth solutions support only the use of wdm as an upgrade for point-to-point networks. In contrast, the immediate pattern of adoption of dwdm is leading to the emergence of photonics technology as a new, valuable logical layer within the network structure. In this architecture, fiber-optic technology is taking on key roles: First was the regenerator, then came the edfa; now the terminal multiplexer with current wdm products is crucial, to be followed by wavelength add/drop multiplexers; and last will be the creation of network restoration within the optical sublayer.
For example, consider the bottom three layers of the International Standards Organization seven-layer description of network functions (see Fig. 1). Note that a new sublayer using photonic technologies emerges at the bottom of Synchronous Optical Network/Synchronous Digital Hierarchy Layer 1. Some network analysts foresee the emergence of this sublayer as a means to rewrite the rules of how public network functions are executed.
In this sublayer approach, fiber-optic technology is being used to create network-level functions, rather than acting "just" as components supporting what is essentially an electronic transmission system. Fiber-optic technology will migrate from the component level to photonic subsystems. Such subsystems provide higher levels of functionality and, hence, value, than do components alone. Consequently, as a business proposition, the value of photonic subsystems scales differently than the traditional optoelectronic components-only business. Of course, the new challenges facing photonics in this network domain are systems-level issues--how to devise new design approaches to lower the cost of wdm multiwavelength technology and how to manage the otherwise burgeoning inventory of components.
In provisioning a 16-wavelength system, for example, all interested parties, from chip makers to field test technicians, must carry or have access to a 16-wavelength inventory of parts. The challenge is to minimize the parts inventory of multichannel optical networks. One solution is represented by a prototype eight-laser transmitter array proposed by scientists from Lucent Technologies` Bell Laboratories (see Fig. 2). Each laser operates at a different wavelength in the wdm grid, but all eight outputs are combined and pass through a single modulator.
The design objective is to create a single laser transmitter that can pass all service and field tests and serve all device functions (see Lightwave, May 1997, page 1). This device epitomizes the future optical challenge--developing fiber-optic component advances that create value in photonic systems. u
John P. Ryan is cofounder and principal at Ryan, Hankin, Kent Inc., a telecommunications market research and consulting firm in San Francisco, CA. He was also the founding editor of Lightwave.