Moving fiber deeper, part two

Moving fiber deeper, part two

By DONALD T. GALL and MITCH SHAPIRO

We continue our focus on the evolution of "fiber-deep" architectures and network development.

While both the cable-television and telephone industries face some greenfield situations, they also face the need to upgrade their existing copper and coaxial networks to support new services. Though very different, legacy telephone and cable networks both place technical and economic constraints on how their owners can respond to demands for new network services.

For example, the cable industry`s distribution network uses a coaxial bus that can deliver relatively large amounts of bandwidth. The size of the "bandwidth pipe" is a strength, especially when the coaxial bus is split into small (e.g., 500-home) sections via fiber-fed nodes. The coaxial bus system has several weaknesses, however.

First, since coaxial cable loses an average of half the signal power every 600 ft, and the bus architecture adds losses at each branch, the network requires amplifiers to extend the network`s reach. These amplifiers need to be robust enough to withstand extremes of temperature and other environmental factors. Cable networks also must contend with the fact that there are a multitude of wireless providers (broadcast television, AM and FM radio, amateur radio, and citizens`-band operators) using the same bandwidth. While the coaxial bus is theoretically a closed system, it is made up of thousands of components in a harsh environment where the odds of being compromised somewhere at any given point of time are high. Yet another major weakness relates to the ability of the coaxial bus architecture to deliver power to devices in the home.

The telephone network, on the other hand, has had more than 100 years to work on the reliability of its distribution network. The twisted-pair delivery system can be very robust and the star architecture from central offices allows network powering of devices in the home relatively easily. This same 100-year heritage is also a major weakness, however. Older telephone plants have twisted-pair cables in service that are many decades old; compromises such as bridged taps and the use of unmatched pairs to bypass trouble make upgrading with electronics to ISDN, ADSL, or VDSL technology expensive and time-consuming.

As for bandwidth, relatively short twisted-pair circuits in good repair, with fewer bridged taps and load coils, can deliver symmetrical data rates above 20 Mbits/sec using VDSL technology. Currently though, much of the older telephone distribution networks in the United States have an effective bandwidth between 33.6 and 56 kbits/sec.

In both the cable and telephone industries, fiber optics is a key to better networks. In cable networks, fiber limits the size of the bus, which allows smaller failure groups. Optical technology also allows smaller failure groups in a telephony plant. In addition, it shortens the twisted-pair paths to the customer, thus opening up more potential bandwidth. While, technically speaking, pushing fiber deeper is almost always a compelling proposition, the case is less clear when economic issues are taken into consideration--as they must be in any business decision.

The process of cost-justification involves an analysis of existing technologies: where they fit into the network, how much they will cost to implement, and any risks they might entail. Hypotheses are then compared to the existing "steady state" business case for validity. In this process of cost-justification there is almost always conflict between the engineers who want to build the battleship and the business elements who are charged with making a profit.

Network cost usually can be broken down into groups of network elements (see Fig. 1). In the majority of networks that serve the general public, the portions of the plant that are not shared typically comprise the largest investment and are the most expensive things to change. For example, a single-customer, plastic interface housing costing $20 is a larger budget item than a $50,000 Synchronous Optical Network router that serves 3000 customers.

FTTC new growth areas

Some local-exchange carriers, including BellSouth and Sprint, have emphasized new-growth areas in their early "fiber- deep" investment programs. BellSouth, for example, has been deploying fiber-to-the-curb (FTTC) in buried new-growth areas since September 1995 and has been escalating the program since then. It currently has FTTC plant reaching more than 200,000 homes and is adding approximately 7000 new homes per month. The company is also planning to begin building aerial FTTC plant soon in new-build situations.

Even in such greenfield applications, however, new techniques for deploying fiber plant have had to be developed to bring initial costs down to an acceptable level. For example, to reduce fiber splicing and craft sensitivity, BellSouth chose to use pre-connectorized multiple-media cables ordered in 150-ft increments. The distribution fiber is built in a star configuration, and collapsed rings are used where the failure group is small enough to warrant them. The current design avoids placing handholes and lays fiber at the same depth as copper, with no concrete markers used for route identification. Care is taken to limit the number of splice points, and there is a higher use of small fiber- and copper-count cables. In front easement situations, BellSouth is building plant down only one side of the street, then serving the opposite side by boring under the street at lot lines to place a pedestal.

BellSouth employs 22-gauge copper pairs using 130V to deliver network power to the optical network unit (ONU). A single pair is used for distances up to 6000 ft from the host digital terminal and two 22-gauge pairs are used to reach up to 12,000 ft. These pairs are contained in the same sheath as the fiber to save cost. From the ONU, five copper pairs are delivered to each house, with copper drops limited to 500 ft. A single fiber carries both upstream and downstream traffic, which provides additional cost savings.

Because it was the first vendor to deliver an FTTC platform that could meet BellSouth`s requirements, Reltec has dominated the telephone company`s FTTC deployments to date, with DSC (now part of Alcatel USA) also approved as an FTTC vendor.

BellSouth is including a video capability in its FTTC platform when it is deployed in areas in which the company operates a video-serving office to support a cable franchise or wireless MMDS operation. This video capability employs simple wavelength-division multiplexer and erbium-doped fiber amplifier technology to carry 750 MHz of video in the 1550-nm window to the ONU. At the ONU, the optical signal is converted to RF, amplified, and delivered to coaxial drop cable using an 8-port tap.

Reltec also offers high-speed data as part of its FTTC solution and has trialed this capability with BellSouth and several other customers. Unlike some other vendors, Reltec delivers high-speed data services from the ONU via a 2-pair Ethernet connection directly to a PC`s Ethernet card. While telephone companies are likely to market this FTTC-based service and DSL-based high-speed data service in a manner that is transparent to customers, the Reltec FTTC approach does not require the relatively expensive DSL modulation components at both ends of the copper lines.

The Reltec approach of delivering "standard" analog video and using Ethernet copper drops for high-speed data contrasts with the approach of other vendors, which appear more focused on ADSL/RADSL for high-speed data or VDSL for both data and video. These approaches typically don`t take fiber as deep as the Reltec FTTC platform and therefore trade off cost savings in fiber deployment with the added cost of DSL gear and, in some cases, a multi-service residential gateway device.

BellSouth`s FTTC initial first costs are believed to be 15-20% higher than those of traditional plant. The company believes, however, that in the long run the reduced maintenance and future service capability offered by an FTTC network will more than justify the increased initial cost.

Upgrading with fiber-deep

Introducing fiber-deep architectures into legacy networks tends to be a more complex decision than a greenfield application, from both a technical and business perspective. Typically, the main drivers for such a step are the prospect of new service revenues to justify the expense or the need to upgrade the network to lower operating costs.

US West`s recently announced Phoenix deployment (see last month`s column) exemplifies the first approach. This strategy seems best suited for high-growth markets like Phoenix that have a lot of fairly new plant, which typically translates into relatively high levels of fiber beyond the central office and a high percentage of "reusable" copper distribution plant. As discussed last month, US West is employing a NextLevel platform that will use VDSL to deliver data rates of 23 to 26 Mbits/sec to support high-speed data and video services delivered via a multi-service residential gateway.

Bell Atlantic`s deployment of narrowband FTTC technology in dense urban areas in Boston, New York, and northern New Jersey is an example of the second, maintenance-driven form of fiber-deep upgrade. A big driver for these rehabilitation projects is the age of the existing plant (in some areas around 100 years old) and the associated maintenance costs. With older network components requiring extra maintenance by relatively expensive urban labor, a compelling case can be made for deploying FTTC technologies, which can greatly reduce trouble call rates (see Fig. 2). This incentive is increased further by regulators ready to impose hefty fines for what they consider unacceptably high trouble call rates.

As noted earlier, the larger the number of customers served by a network element, the smaller its contribution to an upgrade budget will tend to be. This factor is also at play in Bell Atlantic`s FTTC rehabilitation projects, where the average density is more than 600 homes passed per mile of plant. This density not only allows for the network elements to be shared by more customers, it also tends to shorten the individual drops to the customer premises. An overall shorter, more compact plant can more than make up for the higher labor and material costs that exist in the urban market.

BellSouth--already a leader in new-growth FTTC construction--plans to launch a program of FTTC upgrades in 1999, starting in Atlanta and southern Florida. According to a company spokesman, the main drivers for this initiative are high-speed data services--especially in areas not well suited for ADSL--and cost savings in areas with high operating expenses. The prospect of delivering video services may also be a factor in some areas. The company expects the FTTC upgrades to reach 200,000 homes in 1999, with the prospect of doubling that number in 2000.

As we have seen, various combinations of factors are leading different local-exchange carriers (and cable operators) to deploy new, more cost-effective generations of fiber-deep platforms. While we expect this trend to continue and even accelerate, most local networks will continue to face challenging economics when it comes to extending fiber deeper into the "last mile." u

Fig. 2. The prospect of an improved trouble call rate can help justify an FTTC "rehab" for a legacy network.

Donald T. Gall has been involved with the cable-TV industry for the last 28 years. He was an integral part of the team at Time Warner that developed the first practical applications of analog fiber and hfc networks. He is currently a consultant with Pangrac & Associates (Port Aransas, TX) and can be reached at [email protected].

Mitch Shapiro is an independent consultant (and cable modem user) specializing in research and analysis focused on competitive broadband markets and technologies. He can be reached at (760) 753-2890 or via e-mail at mitchshapiro@ home.com.