Multiple fiber-access networks confront network planners with complex choices
Selecting among hybrid fiber/coaxial-cable, fiber- to-the-curb, switched-digital video or fiber-to-the-home network architectures mandates a thorough technical and business analysis to install a future-proof system cost-effectively
paul w. shumate
bellcore
Sorting the technical and cost issues for delivering high-speed data, video and telephony services to homes and businesses proves challenging because the four main fiber-optic networks--hybrid fiber/coaxial-cable, fiber-to-the-curb, switched-digital video (SDV) and
fiber-to-the-home (Ftth)--offer different service capabilities, times to market, cost and revenue profiles, leverages of existing investments, and possibilities of mixing and matching advanced, two-way services. These alternatives, plus one popular in Europe--fiber- to-the-cabinet--are shown in the figure.
Consequently, network providers and planners must carefully study network technology alternatives, capabilities, and technical and business tradeoffs to recognize new application opportunities.
Furthermore, not only are fiber-optic network technologies continually evolving and increasing their capabilities, but customers are also constantly changing their requirements. They are demanding new services, and they are willing to pay for them. These customers, who account for 10% to 25% of all U.S. homes--include Internet users, telecommuters and people operating part-time or full-time home-based businesses. The small-office/home-office market is growing more than 15% annually. Even residences are changing structurally, with home networks and network controller products a growing market.
While analyzing fiber-optics-based networks, network providers and planners should be alert to other service delivery options. The success of digital satellite services and the recent moves by telephone companies into microwave video delivery and updated copper technologies show the dynamic, complex nature of delivering new, advanced and still-unrecognized communications services.
HFC networks
About 20 manufacturers are developing hybrid fiber/coaxial-cable (HFC) networks for delivering data or telephony services to homes and businesses. They are investing notable efforts to address the network problems of ingress, reliability, operations support, and upgradability. The control of noise ingress into the 5- to 42-MH¥upstream (subsplit) band is the leading technical topic under investigation.
Conventional ingress solutions place fixed or addressable blocking filters at the home`s point of entry or use real-time monitoring and spectrum management to relocate upstream carriers away from interfering signals. Recently developed solutions for overcoming noise ingress include discrete multitone (DMT), discrete wavelet multitone (Dwmt), and a spread-spectrum method called code-division multiple access (Cdma). All these methods are resistant to narrowband interfering signals such as those due to amateur, citizens band and other shortwave radio emissions.
Dwmt (developed by DSC Communications) and DMT methods allocate data bits among many (usually hundreds of) transform-related subcarrier frequencies, depending on which ones are free from interference. Spread-spectrum techniques distribute information over a wide frequency range, making them both secure and resistant to interference. Synchronous Cdma (developed by Terayon) allows simultaneous multiaccess operation, a requirement for the HFC return channel.
Other ingress solutions avoid using the subsplit band entirely, at least within the premises where the problem originates, by going to frequencies above 900 MH¥for upstream communications. One approach (developed by Philips) down-converts the premises signals from above 900 MH¥to 5 to 42 MH¥at the point of entry, where local noise can be blocked effectively. The other approach (developed by Lucent Technologies) carries the high-frequency return channels back up through the drop and along the tapped-feeder coaxial cable until they encounter an amplifier (usually located at a distance of less than 1000 feet). At this point, instead of dealing with gigahertz-return amplifiers and diplexers, as well as smaller amplifier spacings, upstream traffic is split off, converted locally to an optical signal, and returned the rest of the way over fiber. This approach results in a fiber-rich HFC network that can later be upgraded by moving the HFC nodes closer to subscribers for more individually dedicated bandwidth.
As to reliability, widely differing analyses exist on how to minimize the primary industry parameter--network unavailability--as measured in average minutes of downtime. Several recent analyses predict an average downtime of 200 to 300 minutes per year. This unavailability is nearly an order of magnitude lower than the value characteristic of cable-TV tree-and-branch networks, but it is much higher than the telephony network objective of only 53 minutes per year (99.99%, which some telephone companies already meet or exceed). Other analyses deduce that some HFC systems are capable of performing better than 53 minutes per year. Obviously, this number is important to operators planning to compete in voice services.
The unavailability differences are due to widely varying assumptions about mean-time-to-repair values; to different architectures, such as no inline coaxial-cable amplifiers versus several amplifiers in cascade; or to the use of widely different values of mean-time-to-failure for components such as amplifiers, power supplies, drop cables, and customer interface electronics.
Because of the growing emphasis on service reliability, many manufacturers have been improving amplifiers and power supplies through redesigns to provide higher electrical surge resistance, better surge protection and tighter screening procedures. In coaxial-cable networks, particularly in the final drop connection, several means are available for improving reliability. These include using special filling or flooding compounds (Times Fiber), incorporating distinct materials in the cable braid to prevent water ingress (Belden), employing more copper (CommScope/GI), improving connector tolerances and installation procedures, and eliminating connectors. All these methods help prevent corrosion, which degrades the cable`s RF performance, particularly at high frequencies.
Beyond the component and media levels, network architectures can be optimized to reduce customer-experienced downtimes to values close to those for fiber-to-the-curb (Fttc) and copper-based "plain old telephone service," although at additional cost.
Lastly, recognizing that cable operators and telephone companies are expecting to use HFC for two-way services that include telephony, and, therefore, need to communicate with existing software systems, several manufacturers have developed comprehensive operations support systems. These software systems provide advanced features for designing, provisioning, monitoring, testing and maintaining HFC networks.
Fiber-to-the-curb
Narrowband Fttc networks have been cost-competitive with traditional copper-based networks in many situations. For broadband applications, several advances have made Fttc competitive. These include using very-high-speed digital subscriber line technology in the drop to permit better sharing of optical network units (that is, obtaining lower costs) in many suburban installations; Asynchronous Transfer Mode (ATM) technology to simplify the switching and multiplexing of video, data and voice services; progress in digital video compression; and optional use of an HFC overlay for electrical powering and analog amplitude-modulated vestigial sideband (Am-vsb) one-way video delivery.
The Am-vsb overlay eliminates the need for an excessive number of digital set-top boxes, which are otherwise needed with all-digital Fttc. The combined system has confusingly been called switched-digital video, which, although capable of providing video service, can also provide other digital services as well as analog video. Recently, network planner interest has shifted back to switched-digital video in the true sense of a switched, digital-access platform, without the analog overlay. Digital satellite services and recent Fttc rollouts have established market viability for digital TV, as well as driving down costs of Mpeg-2 decoders.
Therefore, network operators are increasingly looking ahead to when digital-to-analog conversion becomes inexpensive (even likely to move inside the television set) and to a market preference for high-quality video and audio.
The Fttc/sdv network can be more expensive than HFC if installed to deliver video services first, but it is nominally the same cost as HFC if the initial service is voice. The main reason is that, if telephony is the primary service, HFC loses its cost advantage because every customer then needs an RF interface unit at an additional cost of $400 to $500. Fttc/sdv systems, on the other hand, are designed to provide telephony without additional interfaces. Therefore, while HFC is the clear choice for cable operators and for telephone companies that operate cable systems or install HFC as a video overlay augmenting a conventional telephony plant, Fttc/sdv is competitive where interactivity (telephony and data) is a high priority for customers.
Another important cost factor can influence the choice of network architecture. Industry studies on the annual operating expenses for different networks show that regardless of whether the network is fiber-optic or coaxial cable, the largest cost savings results when the medium terminates at the customer`s premises. Based on life-cycle costs, HFC networks terminating at the home provide the greatest annual payback, that is, until Ftth networks become widely available.
Fiber-to-the-home
Although interest in Ftth networks faded quickly in the United States after the introduction of Fttc to make costs manageable, Ftth development has continued in Japan and Europe. Attention has been focused on network architectures that share costs and on optoelectronics, continually blamed as the culprit for high costs. For example, passive optical networks (which share the equipment at the network end, as well as most of the fiber, among 16 to 32 homes) have emerged from mostly European research but are now embraced by Japan as well.
In the optoelectronics arena, the Bellcore strained-layer, multiquantum-well "loop" laser eliminates the need for thermoelectric cooling and electrical feedback control in the transmitter. Recent innovations in Japan and Europe have increased laser-to-fiber coupling efficiencies while relaxing alignment tolerances, thereby permitting low-cost passive alignment during production. Now, numerous programs are underway to develop planar lightwave circuits to achieve low-cost hybrid integration of transmitter, receiver and wavelength-multiplexing functions on a silicon or ceramic substrate.
Finally, investigative efforts continue to reduce installation costs and clarify how annual operating savings can be obtained from different networks. As a result, the first-cost premium of Ftth over other alternatives has narrowed to perhaps a few hundred dollars (calculated for product quantities greater than 100,000). In fact, in some situations, Ftth can be installed at first-cost parity with other alternatives. Operations cost savings also help to close the gap.
Network power
In the area of powering from network, solutions are available even for Ftth at acceptable cost, as well as low-cost methods that obtain power from the home. Network powering, although more expensive than home powering in terms of both first and annual costs, nevertheless eliminates a roadblock for some service providers: the problem of ensuring reliable lifeline telephone service without having to deal with the issue of maintaining battery power at customers` homes.
However, the power situation is changing. For example, homeowners now accept battery responsibilities for the growing number of cordless and cellular phones, camcorders, smoke detectors, security systems, and uninterruptible power supplies used with home personal computers (PCs).
In addition, battery manufacturers continue to improve their products. Nickel-metal-hydride batteries for PCs and cellular phones minimize the capacity and memory constraints associated with consumer-grade nickel-cadmium batteries, although costs remain high. But new lithium-ion batteries--particularly the type using a "plastic" electrolyte--offer small size, light weight, high reliability, and, for plastic technology, low cost comparable with lead-acid batteries.
Some network providers are promoting the inclusion of uninterruptible power in residential gateway interfaces (see Lightwave, April 1996, page 50). These interfaces interconnect the growing number of home networks (telephone, television, audio/video equipment, PC equipment, security, lighting, etc.) to the growing number of external networks and signal sources (cable TV, telephone company, electric utilities, RF and microwave roof antennas, etc.) to ensure easy and reliable internetworking. Several consumer-electronics industry sectors expect residential gateways and advanced home networks to become a new growth opportunity, helping homeowners cope with the increasingly complex home-electronics environment. If so, residential gateways with uninterruptible power could help accelerate Ftth by finally resolving the age-old question of reliable powering.
For analog compatibility if needed, Am-vsb video can be delivered via an Ftth system either optically (for example, via a 1550-nm wavelength overlay) or electrically, using a coaxial-cable overlay as with an SDV network.
In either case, the Ftth system might be thought of as a second-generation SDV system, with fiber capabilities extended to the home. As the demand for analog video disappears and new powering technologies mature, the coaxial-cable segment is expected to be abandoned; it will likely be timed to coincide approximately with its depreciation life.
The optical overlay is more of a challenge than the coaxial-cable overlay, however. Although analog receivers are inexpensive, the cost of the laser or erbium-doped fiber amplifier behind the final splitter to the customer remains high in the optical overlay; its cost must be reduced to the $500 to $800 range to compete with a coaxial-cable overlay.
Of course, low cost is easier to attain if the number of channels is kept to about 20 or 30; this number permits larger modulation indexes, higher splitting ratios and relaxed laser performance requirements. Because the analog overlay only supplements an extensive digital-video menu, a smaller number of channels might prove adequate.
In sum, Ftth systems are moving along paths that suggest they might be an important factor for full-service access networks by the end of this decade. u
Paul W. Shumate is executive director of broadband local access and premises networks at Bellcore, Morristown, NJ.