As broadband operators work to reduce costs and achieve economies of scale, many are consolidating their local networks into large, regional systems by connecting them with digital fiber rings.

Where each system previously had its own headend, they are now tied into a regional network of one or two primary signal source "super headends" and several distribution hubs. Economies of scale are achieved by consolidating video headend equipment such as scramblers and stereo encoders, as well as the digital switching and routing equipment for high-speed data and voice services.

These regional fiber networks consist of multiple fiber overlays, with individual fibers dedicated for specific services. A typical model consists of at least five fibers for the transport of analog broadcast video (note: uncompressed digital systems support 16 channels per fiber), a second fiber for the delivery of digital broadcast services, and at least two fibers for all interactive services placed on Sonet overlays. Implementing these services in a self-healing, counter-rotating ring doubles this physical fiber route requirement.

With more services being demanded, and competition requiring cost-effective solutions, cable operators face the challenge of adding more services to networks that are already at, or near, capacity. Many are aggressively searching for more bandwidth to make room for additional channels, special programming, telephony and Internet capabilities.

However, creating extra bandwidth can be expensive and time-consuming, particularly if miles of new fiber must be constructed.

A viable solution to this problem is dense wave division multiplexing (DWDM), which enables many information streams to be transmitted over a single fiber. While DWDM technology has existed for several years, recent improvements have made it even more efficient and cost-effective than other options for expanding bandwidth. This article discusses the benefits of DWDM technology for cable operators.

Implementing DWDM

Deploying more fiber has been the primary choice for many years for expanding network capabilities. Each new fiber could add up to 2.4 gigabits per second. This has always been an expensive procedure, but increases in associated costs have made it an even greater investment. First, the average price for deploying new cable is estimated to be as high as $70,000 per mile, and it can be even higher in densely populated areas.

This doesn't include expenses for support systems and electronics.

In recent applications, DWDM has been accepted as a cost-effective alternative when there is not sufficient fiber in place to meet requirements. Many operators have pulled sufficient fiber for their primary headend interconnect routes, but as additional hubs are added to the primary ring, certain segments may contain insufficient fiber.

As a simple cost comparison, Scientific-Atlanta established a hypothetical model to calculate the estimated cost of DWDM vs. the estimated cost of fiber. The modeled regional fiber network had a single headend with up to five subtending hubs, and the hubs were spaced 30 miles apart. In this model, six fibers would be used in the headend-to-hub interconnect. Using DWDM, 16 video channels would be delivered over each OC-48, making the ring capable of transporting either 96 uncompressed analog video channels or 16 QAM IF digital channels and 80 uncompressed analog video channels — all over a single fiber.

Figure 1: Economic advantages of dense wave division multiplexing. As shown in the hypothetical model, the estimated cost of DWDM is favorable not only to the alternative option of constructing new fiber routes, but also to the cost of dedicated fiber for a single headend-to-hub link over 30 route miles, or at the initial hub location. Assumptions: 1. Six fibers are required for the headend-to-hub interconnect. With 16 channels transported in each OC-48, the ring network is capable of transporting either 96 uncompressed analog video channels, or QAM IF digital channels and 80 uncompressed analog video channels. 2. The regional fiber network consists of a single headend, with up to 5 subtending hubs. Each hub is spaced 30 miles apart on the ring network. 3. The model compares the incremental cost of using DWDM with the direct cost of fiber cable. This DWDM cost includes a) ITU wavelength 1550 nm lasers, b) DWDM multiplexer, and c) DWDM demux. Only the cost of the actual fiber cable and splicing is used. No consideration is given to construction costs.

Figure 1 compares the estimated cost of using DWDM in this application with the direct cost of fiber cable. The DWDM estimated cost includes ITU wavelength 1550 nm OC-48 lasers, a DWDM multiplexer and DWDM demux. For fiber cable, only the estimated cost of the cable and splicing is considered. Construction costs are not included.

As anticipated, the model shows that the estimated cost of DWDM is favorable to the estimated cost of the alternative option of constructing new fiber routes. What was not anticipated is the favorable degree of the DWDM interconnect. The DWDM approach for the model is expected to reach cost parity to the cost of dedicated fiber for a single headend-to-hub link at just over 30 route miles, or at the initial hub location. (Note: For routes less than 30 miles, traditional linear 1550 nm transmitters are anticipated.) After the initial interconnect, the expected cost differential is obvious. Estimated savings for DWDM, compared to the cost of fiber, range from 12 percent to 30 percent.

The implications of this simple analysis are significant. Not only is DWDM technology expected to prove cost-effective for the case of insufficient fiber, but another strong business case for DWDM may be realizable even when sufficient fiber does exist.

DWDM application drivers

In addition to the cost savings associated with the utilization of DWDM technology compared to the cost of dedicated fiber, two additional areas of savings must be considered. These are related to the cost of future upgrades and the opportunity costs of fiber.

Future upgrades. Just five years ago in the telco industry, for example, many carriers believed their fiber capacity was sufficient for many years of expansion. However, as Internet and data communications placed huge demands on networks, bandwidths quickly reached capacity. This could very well happen to cable operators. By using DWDM technology, cable operators can multiplex numerous services while saving fiber capacity for future expansion.

It should be less expensive to incorporate DWDM technology during initial system deployment, rather than waiting until bandwidth needs dictate its utilization. An upgrade from dedicated fiber requires a laser transmitter set to a specific 1550 nm wavelength, as well as the DWDM multiplexing and demultiplexing equipment. By postponing the DWDM installation, present fiber capacity may be saturated and the upgrade may require the replacement of costly laser transmitters, not to mention the delay while the new electronics are ordered and installed, and the service disruption during the upgrade.

Installing DWDM in current applications is also expected to be less expensive, because the operator only pays a small initial premium for the transmitters, which should not have to be replaced in the future.

Opportunity costs. While operators often focus on providing core services, such as analog broadcast video and interactive data, they may overlook other opportunities their fiber network may afford. Because cable plant passes many businesses in addition to homes, fiber optic lines can be leased to businesses for LAN-to-LAN networking and new services, including cable modems and even telephony, with little effort. With DS-3 tariffs exceeding $3,000 per month, operators could generate significant additional revenue by leasing fiber capacity to businesses at competitive prices, while maintaining core services using DWDM technology. These revenue opportunities should help offset any incremental costs for initial DWDM implementation.

DWDM capabilities

The use of DWDM technology allows the network to be more powerful and flexible. For example, optical drop-add multiplexers (OADM) can be installed between two end terminals on any route. These enable the operator to drop and/or add up to four STM-16/OC-48 channels between DWDM terminals.

This feature provides the flexibility to allow certain channels to pass through the hub and create express channels. Carriers can tailor services to specific customers and revenue-generating traffic can be distributed, while costs are reduced for installing end terminals at low-traffic areas.

In addition, DWDM systems with an open architecture can greatly assist operators in keeping pace with the constantly changing industry. These systems allow operators to provide Sonet/SDH, asynchronous/PDH, fast Ethernet and ATM traffic on the same fiber. Operators will be able to easily adapt to new technologies without adding optical transmitters for a specific protocol.

Handling the bandwidth crunch

DWDM is the logical choice for the majority of bandwidth traffic jams. New technologies will be able to transport a massive amount of video, telephony and voice traffic over a single fiber. This type of system will multiply the capacity eight or more times, with an anticipated cost savings compared to traditional fiber-optic network architectures.

An 8x DWDM system meshes well with cable operators' current requirements. In one scenario, five optical wavelengths are dedicated to 80 video channels, one for HITS and premium channels, and two for cable modem, telephony and other two-way traffic.

DWDM technology not only lowers initial costs, but also represents efficient revenue generation capability by enabling traditional core services to operate on the same network with two-way DS-3 and Ethernet-based services. When these advantages are combined with the latest optical amplifier technology, DWDM becomes a powerful solution for meeting the bandwidth needs of today's cable networks, as well as providing tremendous capacity for the future.


Author Information
About the author
Tim Wilk is Director of Business Development, Digital Transmission, Scientific-Atlanta Inc.