Hybrid fiber/coaxial-cable networks to expand into interactive global platforms
A three-phase growth path calls for broadband networks to evolve and integrate with worldwide telecommunications infrastructures
andy paff
antec corp.
Fiber-optic technology modernized the cable television industry during the early 1990s and established the hybrid fiber/coaxial-cable network as the preferred platform for delivering an array of video entertainment channels to consumers. Since then, the hybrid platform has emerged as the transport mechanism of choice to deliver other voice, data and multimedia services to consumers.
According to market analyses, the hybrid network provides a scalable solution that portends to deliver a greater bandwidth capacity at a lower capital investment than other technologies. The cable-TV companies, therefore, increased their initial hybrid deployments as increased channel capacity, new revenue streams and improved reliability and signal quality requirements emerged. Hybrid platforms can also deliver digital signals, in various digital modulation formats, and enable broadband operators to migrate gracefully to the all-digital transport network expected in the future.
The migration prospect has prompted the attention of telephone companies that seek to expand their own network capacity to the home. Today, most hybrid systems are positioned at 750 megahert¥and can deliver 80 analog video channels in the 50- to 550-MH¥range with 200 MH¥reserved for digital services. Non-active components, including broadband taps, splitters and connectors, can accommodate a total bandwidth of 1 gigahertz.
The return path
The traditional return path lies in the 5- to 40-MH¥spectrum, though additional return spectrum can be utilized between 750 MH¥and 1 GH¥if all radio-frequency amplifiers, which cause feedback and degradation to return signals, are eliminated from the network.
To fully position the hybrid network for interactive capabilities, the size of the areas served by fiber-optic cable remains a critical factor. Such interactive services as broadband telephony and data communications rapidly absorb return-path capacity, which is currently limited to approximately 25 MH¥in the 5- to 40-MH¥range.
With only 25 MH¥available until the coaxial-cable portion of the hybrid network becomes fully passive (by eliminating all RF amplifiers), hybrid systems that size their optical nodes to support 2000 homes will allow only approximately 25% of the customers to simultaneously use their telephones; for example, before the hybrid network runs out of capacity. For optical nodes sized to serve 500 homes, almost 96% of the customers would have simultaneous telephone use before the network would be overextended.
Most advanced hybrid designs size their service areas to 500-home nodes in anticipation of this type of capacity need, especially in the return path.
Forward-thinking cable operators are also pre-provisioning 500-home serving nodes with additional fiber to eventually segment these serving areas down to less than 100-home "passive" nodes to eliminate active devices. Then, the system`s high-end return capacity (750 MH¥to 1 GHz) could be used for additional return information.
A three-phase approach
Considerable progress has been made in hybrid network evolution, but more studies remain to tie the residential platform to the wider global communications infrastructure. One such study involves a three-phase approach to hybrid network growth that provides a cost-effective and practical migration plan.
Phase 1 deals with regionally interconnected headends, the signal origination point for cable-TV operators, and provides key benefits for the core application--analog video delivery.
These benefits include:
Achieving operational economies by consolidating operations at a single site
Enhancing the network`s advertising business
Delivering enhanced pay-per-view (analog video on demand).
Regionalization of existing hybrid networks achieves operational economies by eliminating many business locations that typically serve small groups of subscribers (for example, 60,000 homes). By centralizing equipment and personnel resources, one master site could serve 200,000-home geographic areas more efficiently than headends could if deployed over that geographic area. Once interconnected, previous headend locations would act as remote signal-processing hubs.
A double-star network design establishes an ideal residential foundation. In the residential component of this network, optical transition node sites initially function as optical repeater sites dee¥in the residential plant. These nodes are optimally sized to serve no more than 20,000 homes and initially receive, split and process broadcast hybrid signals to feed no more than 40 optical nodes of 500 homes each.
Because this first phase involves forward-only broadcast video, fiber counts to the optical transition nodes are reduced to cut costs in the residential plant. The nodes also spread headend processing equipment costs, such as RF modulators, over a larger subscriber base.
Operational economies are thus achieved through the management of one network rather than several. The master regional headend would acquire all the signals and process them to the remote hubs (formerly headends), which then feed the nodes.
Moreover, this setu¥boosts the hybrid advertising business by providing an integrated processing and distribution center for delivering commercials to consumers.
New digital advertising insertion systems allow vendors to target their messages to demographic groups and operators and to sell more advertising time. Digital advertisement insertion could start by targeting one ad per region (200,000 homes), then move to target ads to each remote hub (60,000 homes), to one ad per node (20,000 homes), and finally, to one ad per optical node (500 or fewer homes).
In this manner, like the hybrid network itself, advertising insertion could grow in a scalable, upgradable fashion as new revenue streams from advertising are built.
Video traffic
Trafficking video through the region may initially involve proprietary digital technologies, though ultimately, the synchronous optical network will become critical to the regional network in moving voice, data and digital video applications and internal telecommunications traffic.
Full Sonet integration at the regional level allows the transfer of information to and from the public network--a critical component in delivering interactive and transactional services. Also during this phase, the hybrid network is connected to long-distance providers to offer global access.
The digital compression standard, motion picture experts group-2, is commonly used to digitize and compress entertainment video signals, and could prove the impetus for rapid Sonet deployment. With this standard, MPEG-2 cells can ma¥directly into the Sonet synchronous transport signal, level n frame or can be converted to asynchronous transfer mode cells when further processing is needed.
At remote headends, entertainment video could be decoded or, when possible, transmitted in compressed form through the hybrid network for decoding at the consumer`s set-to¥device.
Hybrid network providers should carefully chart the assumed compatibility among the Sonet, ATM and MPEG-2 standards. The compatibility issues have not yet been resolved, but various forums and committees are working on the important interoperability aspects.
Phase 2 integration
Phase 2 is projected to accommodate prospective digital applications. These applications, which will require dedicated bandwidth to and from the home, include video-on-demand for both advertising insertion and MPEG-2 delivery to the home set-top; such data and multimedia communications as online service connectivity and teleconferencing; and broadband telephony.
An important element of Phase 2 will be the implementation of a common transport standard to optimize hybrid network use. Several transport techniques are under investigation. ATM cells modulated in 64-quadrature amplitude modulation format, for example, could be used for downstream traffic (headend to home). In another method, ATM cells modulated in quadrature phase shift keying format with some multiple access protocol, such as time-division multiple access, could provide the upstream (home to headend) support.
The actual implementation, particularly in the initial part of Phase 2, is expected to be complex. A variety of formats and protocols will probably prove practical to transport a particular grou¥of applications and manage the limited return path.
Although one strength of the hybrid platform is its ability to handle different formats through discrete frequency assignments, as much commonality as possible will drive a standardized digital transport format.
As networks move into Phase 2, dedicated lasers should be available to segment the 20,000 homes into individual nodes. This segmentation provides the ability to "narrowcast" services--the delivery of a specific application only to a given node. Low-power lasers (3 to 5 milliwatts) would cut capital costs in establishing narrowcasting where a single laser feeds a single optical receiver. Optical receivers, too, would be retrofitted with return lasers, or installed with them.
Initially, upstream frequency block converters at each node could be used to stack the incoming 5- to 40-MH¥return information from each node (Node A handles 5 to 40 MHz; Node B, 52 to 88 MHz, etc.) for transport via a separate optical fiber back through the regional network for processing.
During this phase, the optical transition node should be able to receive bidirectional signals. Because interactive applications, including telephony and data communications, will involve the public network, the Sonet backbone network becomes the logical choice for linking nodes. Compressed digital video applications, whether in native MPEG-2 or ATM-cell format, can be delivered to the nodes through Sonet optical carrier, level 3 (155-megabit-per-second) or OC-12 (622-Mbit/sec) rings. The master headend and remote hubs could be linked via an OC-48 (2.5-gigabit-per-second) regional ring.
Video file servers would also move into the optical transition nodes. Storage and processing for advertising insertion, analog video-on-demand (broadcast), video-on-demand (real-time, dedicated channel), catalog shopping and games would move closer to the customer. The nodes are also likely to house some routing functions, including an ATM switch.
With the Phase 2 requirements for digital modulators, dedicated lasers, Sonet equipment, video file servers, routers and ATM switches, the optical transition nodes would physically expand into a unit somewhat larger than an environmentally controlled roadside cabinet. Because additional space and local regulatory approvals might be required, initial node placement must be made with future services in mind.
Broadcast modulators are likely to remain at the remote headend. However, changes are forecast in digital grooming and network management. For digital video requirements alone, the Sonet backbone network would have to expand rapidly. Each 51.84-Mbit/sec OC-1 (STS-1) frame will carry approximately seven National Television Standards Committee video channels, or perhaps even 10, given the dynamic compression capabilities of MPEG-2.
However efficient MPEG-2 becomes, routing dedicated video material to individual nodes will consume Sonet bandwidth on the backbone OC-48 network and the OC-3/OC-12 node rings. Crossconnects and ATM routers might eventually be located at the remote headend for optimizing Sonet bandwidth.
Remote headends should also serve as primary points for the hybrid network element manager. Although the hybrid element manager still remains generally undefined, it will need to monitor and control such elements as modulators, lasers, receivers, power supplies and all active devices past the optical node.
Phase 3 integration
After service information is transported to and from homes in digital format, the need for additional digital bandwidth becomes critical. Analog broadcast modulators would disappear, and digital modulation would become typical. During this phase, the optical transition node would contain all the RF modulators, assuming equipment cost reductions follow volume increases and size is reduced.
In the expected all-digital world, the conventional sub-split configuration (5 MH¥to 40 MH¥and 750 MH¥to 1 GH¥for the return path, and 50 MH¥to 750 MH¥for the forward path) could be addressed. In addition, upstream/ downstream signal requirements could be readily balanced, given node-level demand. Fiber should move closer to the home, and the coaxial-cable bus should become passive as all remaining RF amplifiers are eliminated from the hybrid network. The high-end spectrum (750 MH¥to 1 GHz) would be used for return information without the conventional feedback problems caused by the RF amplifiers.
The Sonet bandwidth in both the backbone and node networks should remain crucial. Distributed switching, routing and grooming at remote headends need to be economically compared to deploying this equipment at the node to increase Sonet bandwidth. Once OC-96 (5-Gbit/sec) products are available, the tradeoff between ring capacity and the location of processing equipment would need to be balanced. u
Andy Paff is executive vice president for strategic planning and technology at Antec Corp. in Englewood, CO.