Digital reverse solves the upstream bandwidth crunch

As cable operators upgrade their networks, they are looking for better ways to use existing equipment, control costs, and still increase network capacity. This need is especially evident in the implementation of upstream transmission of bandwidth-consuming, interactive multimedia traffic.

The deployment of the reverse path has brought about several major challenges for cable operators. To deliver new services such as multimedia, Internet access, and telephony, and increase subscriber penetration, operators must "enlarge their pipes" and limit the number of users served by a given node. Operators also need to allow more traffic on a single fiber while retaining their existing architectures--if possible--to control costs. With existing analog technology, however, cable operators face a dilemma: find more bandwidth or serve fewer subscribers.

An emerging solution is the application of baseband digital technology in the reverse path. While the use of baseband digital has been technically possible for some time, the associated analog-to-digital conversion technology has only recently developed to the level of environmental hardening, speed, and cost needed by cable operators.

Existing problems - The structure of existing networks makes implementing an effective reverse path daunting. These networks were originally designed to transmit from one source to many locations and now must be modified to also transmit in the reverse direction-from many sources to one location.

With a greater number of signals to process, and because the analog signal is subject to losses and distortions directly proportional to distance, many operators have been forced to configure their networks to process the upstream traffic at the hub. This action may involve receiving, demultiplexing, frequency conversion, re-modulation, and retransmission of reverse signals received from multiple node locations. By necessity, hubs have evolved to become larger structures, often requiring additional rights-of-way, staff, power, and air conditioning, making them expensive facilities to operate.

To transmit the signal from the hub to the headend, some operators have implemented radio-frequency (RF) combining, block conversion, or reverse dense wavelength-division multiplexing (DWDM). These solutions work but quickly become complex and expensive when node sizes are reduced and the number of upstream paths increases. The cost of implementing reverse DWDM for a fiber-optic node return fiber can cost up to twice as much as the node itself.

Reverse-path traffic-management problems are exacerbated in larger cable networks and in networks serving high-density areas where more taps and active components are deployed per mile. The likelihood of reverse-path problems, such as ingress increases, makes isolation of such problems a tremendous challenge. Moreover, because of headend consolidation and a regional service structure, the distance between the headend and the customer's home is often increased to a point where it becomes difficult to ensure a reliable signal is transmitted, especially in the upstream direction. In this enigmatic situation, the need for additional bandwidth as well as to limit node serving sizes forces operators to push nodes and headend equipment closer to the home. Maintenance costs and capital investments for the same equipment, however, drive it away from the home.

Fig. 1. Analog reverse-path performance is often unpredictable due to the effects of noise and temperature, which limit data capacity and quality. It is also impractical and costly to combine and transport analog signals at the hub. Within the traditional analog network architecture, the performance of an upstream signal is often unpredictable, suffering from the effects of noise and temperature between the node and the hub (see Fig. 1). For low-bandwidth applications, this transmission is often handled by Fabry-Perot lasers, which are inexpensive but limited in handling large volumes of multimedia traffic. For higher-bandwidth applications, uncooled distributed-feedback (DFB) lasers offer better performance but at a higher cost.

Comparatively, the cost of using baseband digital reverse transmission is now much more reasonable. This transmission technology can actually lower the comparative cost of upgrades while introducing new services to more subscribers.

By riding the cost curves and applying technology originally developed and used for digital two-way transmission in the telecommunications industry, cable networks can use digital capability to carry high volumes of multimedia traffic over longer distances, at a lower cost, than with comparable analog solutions. With its increased capacity, the network is able to serve more subscribers on a single fiber with a robust signal that is not prone to the problems associated with analog transmission.

How it works - Cable operators can greatly increase reverse bandwidth by using time-division multiplexing (TDM) at the node and installing digital transmitters. This setup allows operators to convert analog signals to digital at the node, then back to analog with receivers at the hub or headend. In effect, this architecture solves two problems: how to reduce the return fiber count by using more return bandwidth per fiber and how to transmit a return signal from the hub to the headend at lower cost with better end-of-line performance.

Fig. 2. Digital technology provides better performance, including delivery over long distances, greater dynamic range, and lower costs. Data processing can be centralized at the headend, with simplified transport capabilities, eliminating the need for extensive processing at the hub. Because of the robustness and reliability of the digital signal, it can be transmitted from the node all the way to the headend, creating the option of an entirely passive hub architecture where only passive optical combining is needed. Although node cost is higher in this design, it is more than offset by lower costs at the hub. This design eliminates the need for receivers, transmitters, amplifiers, and associated processing equipment. It also allows operators to reduce the physical hub size as well as the associated maintenance and overhead costs (see Fig. 2).

Some operators prefer to use their existing hub architecture, while others choose to build networks with smaller hubs or no hubs at all. Digital technology has the flexibility to meet either need, allowing operators to select where they want to process the information. Operators can often retain much of their current architecture by simply plugging in digital transmitters and receivers instead of analog devices.

Baseband digital reverse technology also makes the network easier to manage, with opportunities for digital signal processing (DSP) functions such as filtering, noise suppression, and data manipulation to enhance performance. Because the digital signal is relatively immune to the effects of noise and temperature, it offers more reliable performance over long distances. It also enables greater bandwidth exploitation than analog, which is critical for delivery of two-way multimedia traffic. The added capacity, enabled through TDM, allows operators to serve a greater number of subscribers on a single fiber.

Lowering costs - DSP costs have decreased in recent months, and prices are likely to continue to fall in the near future. And with digital technology, a single fiber supports many more subscribers.

A digital signal reduces maintenance costs because transmission is less likely to be affected by noise, temperature, frequency roll-offs, and other analog aberrations. For every decibel of change in optical level in an analog system, the RF-level and performance are correspondingly affected. The performance and RF output level of the digital reverse system, however, remains constant even as the input level is varied over 20 dB. This has a very positive impact on setup and maintenance of a digital reverse system. Subscribers' satisfaction is increased because the signal is less likely to be interrupted for maintenance. And the option of bypassing the hub and transmitting directly from the node to the headend lowers expenses for operators in the process of building networks or creating passive hubs.

Qualities inherent in the digital reverse signal allow many components, such as demodulators and routers, to be moved upstream to the headend-or to be entirely eliminated--creating economies of scale at the headend and simplifying maintenance overall. This capability also lowers costs by eliminating the need for large, expensive hub facilities.

Since 1996, the cost of digital technology, including lasers and DSP, has become increasingly smaller, with much lower costs per milliwatt. Digital technology is now the logical choice when compared to other technologies, such as Fabry-Perot transmitters or DFB lasers. If Moore's Law is any indication, this trend will continue over the next few years.

Ideal applications - The flexibility and power of digital technology make it suitable for a variety of situations. In existing networks, operators can install receivers at the hub or headend, depending on preference, architecture, and constraints for where the information is processed. During the construction of a new network, operators can dramatically lower hub costs by designing the network for transmission directly from the node to the headend, eliminating the need for processing equipment at the hub. The digital reverse path is viable for densely populated areas because its exceptional capacity allows a single fiber to support a large number of subscribers.

Digital technology allows operators to place more information on a fiber while increasing network throughput, reliability, and signal quality. The high bandwidth capacity provides the ability to transmit a range of services in the reverse path, including cable modem traffic, telephone services, and other interactive content. It also provides the extra bandwidth necessary for future network expansions.

Virtually every industry is realizing the advantages of digital solutions, and in the next few years, almost any product that can be converted to digital will undergo the switch. With the introduction of an array of new interactive services and ever-expanding channel lineups, the cable industry is certainly a prime example of a market that will benefit from digital technology.

The complex issues of reverse-path transmission make digital especially viable to cable operators. The need to cost-effectively deliver numerous signals to a single source, carry high volumes of multiservice traffic, and improve the reliability and management of the network, make digital technology the logical choice for robust cable networks. As cable operators attempt to meet the demands of subscribers and stay ahead of growing competition, digital is the logical solution to reverse-path problems and the implementation of new services.

Paul Connolly is vice president of marketing and network architectures for Transmission Network Systems at Scientific-Atlanta Inc. (Lawrenceville, GA).

More in Network Design