Optical Communications Markets' Hidden Strength in Metropolitan Area Networks by Dr. Scott Moore, Director of Technical Assessment The apparent collapse of the telecommunications industry, combined with a significant slowdown in other bellwether technologies like semiconductors, hides a significant strength in an important part of the technological landscape: photonics components and systems developed for communications. Many of the telecommunications industry's troubles can be traced to the dot com hysteria of recent years, a time when venture capitalists were eager to throw enormous amounts of money at questionable Internet-based startups - including capital-intensive telecomms plays. The frenzy gave rise to the notion that these new e-businesses required a tremendous amount of hardware - mostly servers - to support their efforts to move the world's economy to a form where many transactions are performed over the Internet, often replacing "brick and mortar" stores. Deployment of so much infrastructure caused the telecommunications industry to gear itself up to accommodate the expected increase in Internet transactions. The primary focus of this effort was to lay optical fiber, the backbone of modern telecommunications, where voice and data are sent over networks in the form of light rather than electricity. The tremendous information-carrying capacity of optical fiber, referred to as bandwidth, where a single fiber can easily carry thousands of voice conversations, forms most of the telecommunications infrastructure in the more advanced countries. Even wireless communications relies to a great extent on optical fiber networks. The deployed optical fiber was largely for long haul networks, those spans of fiber approaching or exceeding a thousand kilometers. Placement of long haul fiber networks is in many ways an easier, less complex task than in other types of networks. Carriers made the decision to emphasize long haul fiber deployments based on projected growth in the dot com boom, rather than other sectors where the need for increased capacity was more acute. The bust in the dot com boom, coupled with some spectacular errors in business judgment (as with Lucent Technologies), led the media (and many investors) leapt to pronounce a "bandwidth glut", a far greater capacity in place than could be used. As prices for network access fell, the notion that the optical infrastructure had been overbuilt became the conventional wisdom. The reality is a little more complicated. The overbuilding in long haul networks resulted from projected growth in a specific sector of business: the newer Internet and e-commerce companies. Had the vision of telecommunications carriers been a little more clear, the building boom may have targeted a more useful segment, one where the tremendously increased capacity could have been used to fill existing demand rather than lay idle: namely, local metropolitan and access networks. The demand is there, has been there and will continue to grow. While many of the new dot com companies have not attracted the level of business first expected, the end user - businesses and consumers using the Internet for business and pleasure - continues to demand more capacity at lower prices. Basic market drivers may have softened only slightly but are still driving a remarkable surge in demand for bandwidth, the basic measure of the amount of information-carrying capacity of networks. The failure of a number of companies based on an untenable business model has masked the extraordinary demand that brought about those companies in the first place. Already, data transmission over telecommunications networks exceeds voice communications. According to Insight Research, Internet traffic is growing in excess of 300 percent per year. On the transatlantic corridor, Internet traffic surpassed voice communications in 1997 and exceeds 100 percent growth. Research firm Insight projects worldwide Internet traffic to grow to more than 800,000 Gigabits per second by 2006, compared to less than 5,000 Gigabits per second in 2000. Internet traffic growth is driven by an increased number of users, a greater variety of services available via the Internet, and more demand for those services. Email and the World Wide Web, used now almost universally by businesses, are significant sources of Internet usage, used for business, entertainment, research and telecommuting. E-commerce, both business to business and business to consumer, have worked to transform the world's economy. The number of online shoppers has increased at an even greater rate than that of overall Internet users, to the point where soon most users - which will soon mean most people - will regularly use the Internet for purchasing goods and services, more than will use catalogs and shopping by telephone. Greater capacity in telecommunications networks will be required to handle growing demand. While it is true that on long haul networks much capacity is idle and falling in price, the primary reason that the networks are "dark", or unused, is because much of the communications traffic is stuck in smaller, more local networks, unable to access the long haul optical backbone. The situation is much like having an enormous traffic jam in a metropolitan area, with a few small access ramps to the highways choking off so much traffic that the surrounding highways are nearly empty of traffic. It is, in fact, the so-called metropolitan and access networks (MAN), those where fiber spans are typically 20 km between network nodes, that users rely on to access the high speed long haul networks. The rapid growth in data communications (such as Internet access) has resulted in a tremendous strain on metro networks; if the congestion could be alleviated, much of the idle capacity in long haul networks would be used more efficiently as traffic passes the local bottlenecks. Thus, the emphasis on overcapacity has been largely misplaced. Rather, there are local bottlenecks at the metropolitan and access network level that must be relieved. If anything, there is an overall undercapacity, relative to bandwidth demand by the actual users, currently and in the next few years. MANs present some unique technical challenges that must first be solved. Many are based on an increasingly obsolete technology referred to as Synchronous Optical NETworking (SONET). Developed in the 1980s, SONET uses costly and inflexible technology with limited communications-carrying capacity that has offered diminishing gains as performance requirements have increased. Long haul networks, on the other hand, mostly make use of a more modern approach: DWDM, or Dense Wavelength Division Multiplexing. DWDM allows the sending of several different wavelengths of light, or channels, over the same optical fiber, allowing for greater capacity. Thus a single fiber can carry many channels, for a bit rate that is effectively many times higher than a fiber carrying only one. DWDM systems used in long haul communications are complex, sophisticated and expensive. For a metropolitan area network, they are serious overkill, too costly to provide a realistic solution to local congestion problems. To bring the advantages of DWDM to the local level, systems must be developed that will offer somewhat reduced capacity - and therefore reduced cost - while solving some other problems unique to local networks, like a need for flexibility in supporting multiple architectures, making capacity adjustments based on instantaneous traffic levels, and interoperability with older equipment. A major weakness of today's DWDM equipment is lack of flexibility. The problem of congestion in MANs will require significant advances in the technology of optical communications. Even in long haul networks, scientists are continuing to develop technology that increases speed and capacity of networks. The holy grail of fiber optic communications is the all-optical network. Today, as data is relayed through a network, it must be converted from optical to electrical signals at the various nodes to be processed and routed, then converted once again to an optical signal and sent on its way. The conversion from light to electricity and back to light slows the whole network considerably, and adds to the network's overall cost. Companies in the photonics business are making intense efforts to develop components for the so-called all-optical network, one where data can be routed throughout the network, from start to finish, without any conversion to electricity. When developed, the all-optical network promises to not only speed up communications tremendously and relieve bottlenecks, but to increase efficiency to such an extent that network access costs will be considerably cheaper than today. In 2001, several companies are marketing systems that perform network control tasks optically, without conversion of signals. There remain technical obstacles in the way of full development of all the systems required in telecommunications - switches, routers, multiplexers/demultiplexers, cross-connects, etc. - with all-optical technology. The existing technologies designed to redirect light in different directions in such a system, a basic requirement for directing packets of information to a destination, remain primitive and inefficient. For at least the next five years, most of the major routers and switches will still make use of optical to electronic to optical conversion. In the meantime, system designers are working to reduce the number of electrical interfaces in optical networks through more efficient design. There are also fundamental issues that must be addressed before all-optical devices can be widely incorporated in the communications infrastructure. Industry participants are still debating whether all-optical network elements as envisioned will scale to the tremendous sizes required by carriers, and whether the technology can be developed in a way that it will be able to manage networks in the total absence of electronic interfaces. The largest all-optical cross-connect on the market has a 256 x 256 port capacity, compared to the 1024 x 1024 capacity needed by carriers. On top of the other concerns, any all-optical network subsystems must offer significant cost savings if they are to make any headway at all. If the equipment cannot be manufactured inexpensively, it won't be used. Thus the continued advance of the all-optical network depends not only on development of high performance technology, but technology that competes economically with that of established systems. The ongoing congestion in local telecommunications networks and the extraordinary increase in bandwidth demand will likely work together to force solutions that will continue to make more abundant and ever-cheaper communications services universally available. Certainly, the demand will continue.

US Offices: 38 Rice Street Suites 2-0/2-2 Cambridge, Massachusetts, USA 02140 VOX: 1-617-491-2952     FAX: 1-815-364-3002 Email: info@triarche.com | Genesis Myth | Consultants | Commentary | Partners | International Bureaus | Network Architecture | Marketing Campaigns | E-Commerce Stategies | Research | Technical Assessment | Homepage | Contact Us | All Rights reserved: Triarche Research Group Copyright: 2000