Single-fiber approach cuts cost in last mile

For several years, the fundamental inhibitor to the deployment of Ethernet services has been the availability of fiber. While fiber is currently available in many downtown locations, it is generally less available in suburban office parks and other customer locations. In the boom times, carriers could-and did-aggressively extend the fiber footprint of their networks. Given the general turmoil within the service provider community, it is fair to say that today's economic climate has put many proposed fiber buildouts on hold, effectively capping the rollout of optical Ethernet services to the current network footprint. But recent developments in the fiber transceiver market have greatly advanced a carrier's ability to offer Ethernet services where fiber scarcity is a problem, allowing them to target more customers for high-value data services without the cost of trenching new fiber.

Historically, two-fiber transceiver designs have dominated applications with fiber endpoints, such as Sonet add/drop multiplexers, Ethernet switches and point-to-point fiber links. Now single-fiber designs are emerging to challenge the dominance of two-fiber designs, especially in the last mile of the telecommunications access network. The key advantage of single-fiber transceivers is that they conserve fiber, doubling the number of locations or devices that a carrier can serve with existing infrastructure. These transceivers potentially halve capital equipment costs, use space more efficiently and lower operating expenses.

Today, a vast majority of all fiber installations employ two-fiber transceivers. One fiber transmits the downstream data path and the second fiber transmits the upstream direction. Two-fiber designs are sufficient for traditional fiber communications markets, including both the metro-backhaul and long-haul networks.

In these metro and long-haul applications, the fiber infrastructure is laid over long distances. In many cases there is sufficient dark fiber to handle the increasing demand for bandwidth for some time to come, and the market is not particularly cost-sensitive.

Single-fiber transceivers use a single strand of fiber to simultaneously transmit in both directions, effectively doubling the capacity of fiber investment. But due to a lack of standardization, these transceivers have been used only where absolutely necessary in niche or special applications such as the military, for example.

Today, however, conditions are ripe for an explosion in the use of single-fiber transceivers, and component vendors and standards bodies are actively supporting this move. The short distance, high volume and cost sensitivity of the access market lend themselves particularly well to single-fiber, bidirectional solutions. A primary driver for migrating to single-fiber access is its favorable economics, the result of the volume deployment of access systems and the cost of deploying the fiber.

Volume deployment of access systems provides the impetus to drive down the cost of single-fiber transceivers. This catalyst began several years ago in Asia, when carriers started deploying fiber to the home (FTTH) and offering 100-Mbit/second (Fast Ethernet) speeds. To most efficiently use the limited amount of already installed fiber and to minimize the footprint of access equipment in space-constrained high-rise apartments, customers started pushing carriers to do FTTH on a single fiber and invested money to help solve the problem. As a result, single-fiber transceivers in Asia are now on cost parity with dual-fiber transceivers at the Fast Ethernet data rate. These devices are available from a number of vendors in a variety of package styles and connector options.

For mass deployment in the access market, however, single-fiber transceivers need to be standardized. The Japanese Telecommunications Technical Committee has developed its TS-1000 standard addressing the single-fiber interface for Fast Ethernet. The IEEE's Ethernet in the First Mile (EFM) working group is currently developing single-fiber specifications for 100-Mbit/second Fast Ethernet and Gigabit Ethernet applications. The feasibility stage is progressing and the IEEE 802.3ah standard should be ratified early next year, which gives manufacturers plenty of time to tool their designs. In the interim, some manufacturers are working on early deployments based on existing technology and designed to accommodate the standards that are now in process. This standardization effort will allow equipment manufacturers to multisource components and further reduce cost.

Another single-fiber area where standardization is progressing is in passive optical networks. The ITU has had G.983 available for several years while the 802.3ah standard addresses an Ethernet version of this architecture.

The worldwide growth of Ethernet services over the next several years will drive volume deployment for equipment and transceiver manufacturers, and thus lower the cost of single-fiber solutions. Industry analyst group IDC predicts the number of metro Ethernet service subscriptions will grow from roughly 1.2 million in 2002 to 18.7 million by 2006 worldwide, with a compound annual growth rate (CAGR) of 120.1 percent That figure includes companies that have more than one Ethernet service per company. More importantly, IDC also reports that new Ethernet connections-meaning new Ethernet service customers-more than doubled in 2002, indicating the growth potential of this market. IDC forecasts that new subscriber connections will grow to 6.9 million by 2006, with a CAGR of 67.5 percent.

The second compelling factor for developing single-fiber transceivers is the cost of fiber deployment. The actual cost of cable and of trenching or aerial mounting is relatively insensitive to the number of fibers in the cable. However, the cost of splicing and components such as splitters, connectors, couplers and splice enclosures is directly related to the fiber count. This is especially important in access applications, since these tend to be "splice-rich" as the fiber branches out from the large central-office cable to the single drop at the customer end.

These costs tend to add up quickly in access networks, where subscriber demand lies. Single-fiber solutions help reduce the capital and labor costs associated with lighting up fiber while effectively doubling the number of available access fibers and the number of customers a carrier can reach.

Single-fiber solutions also maximize the capital investment in new fiber installations and make better use of existing dark fiber by extending the lifetime of fiber and thus extending "time to trench." Metro-area networks nearing fiber exhaust also represent a potential market for single-fiber transceivers, which can turn 10 percent fiber availability into 20 percent availability.

Single-fiber, bidirectional technology is fast emerging as the low-cost solution for the Fast Ethernet access market, where demand is driven by the need for fast Internet access.

In the future, the same technology can change the cost model for Gigabit Ethernet access as well. Carriers that install a hybrid copper/single-fiber transceiver solution can provide 1- through 100-Mbit/s access service to customers today and be positioned to support Gigabit Ethernet bandwidth when the demand arises.

Mark L. Ventura is vice president of hardware engineering for Hatteras Networks (Research Triangle Park, N.C.).

http://www.eet.com