Filling up the fiber with digital TV
SPECIAL REPORTS / Access Networks
As technologies come together to allow a new revenue opportunity for telecom carriers, what is the best technical strategy to capture a portion of the home-video entertainment market?
DAVE PECORELLA and MARTY DUGAN, Artel Video Systems
As cable television operators move to bundle video, voice, and data offerings and talk of launching video-on-demand and interactive content services, telecommunications companies must form competitive service strategies. The missing link in the telecom carriers' service bundle is video-a huge revenue opportunity worldwide.
How will video be delivered to the home over phone lines? It's clear that DSL technology will be used as a delivery infrastructure. At the end of 2000, the number of DSL data subscribers worldwide topped 5.8 million. That number is expected to surpass 150 million by 2005, with 60% of the subscribers outside of the United States.
If this same infrastructure can be used to deliver end-to-end video services to those subscribers, what is the best technical strategy to capture a portion of the home-video entertainment market? Several technologies are coming together to enable full-screen digital television delivered over DSL. Powerful MPEG-2 and MPEG-4 video compression algorithms, coupled with intelligent ATM video-aware network edge devices, allow service providers to leverage their existing fiber-optic core ATM network.
The growth of DSL will promote deeper fiber penetration. Provisioning digital TV services over unshielded twisted-pair (UTP) telecom networks requires transport of the video data over core fiber infrastructure from centralized play-out centers to regional central offices. These DSL provisioning points are not merely at the edge of the metropolitan or wide-area network but deeper into neighborhoods, streets, multidwelling units, and office buildings. The requirement of deep penetration of fiber is directly related to the distance and bandwidth limitations of DSL technology. To adequately deliver digital TV services to subscribers requires 4-6 Mbits/sec of bandwidth to the home. DSL can only provide such bandwidth at distances of less than 8,000 ft from a central office. Therefore, deep fiber will play a critical role in the deployment of digital TV services as DSL access multiplexers (DSLAMs) move into the neighborhoods.
Traditional UTP local loops and the central offices that terminated them were designed to provide voice-only services. DSL is an access technology deployed in "last mile" local-access networks. That relatively non-intensive bandwidth application has the advantage of not being particularly sensitive to the two things that degrade a local loop: crosstalk and attenuation.
Asynchronous DSL (ADSL) provides enough bandwidth to provision video, data, and voice services. Using discrete multitude (DMT) modulation, a technique developed by AT&T two decades ago, downstream data rates can be provisioned up to 6 Mbits/sec with an upstream path of up to 1 Mbit/sec. There are, however, limitations on the length of the carrier loops as with any high-data-rate UTP-based transport. To achieve these bandwidths, the data path for ADSL can only be about 6,000 ft. with good line quality, which is achieved by clearing unused taps and loading coils. Very-high-data-rate DSL (VDSL) is another option, which allows for up to 52-Mbit/sec data rates at 1,000 ft. from a central-office service point. VDSL is still distance-limited but would allow for 13 4-Mbit/sec video channels of service.
The compelling reason to use ATM for video services is that most of the existing voice traffic in the core of the public-switched telephone network (PSTN) uses ATM/SONET as a transport medium. The data traffic from multiple loops gets concentrated on a DSLAM. DSLAM devices are placed in the local central office, and sometimes even at the curb, with fiber run into the neighborhood carrying the ATM traffic on an OC-3 (155-Mbit/sec) or OC-12 (622-Mbit/sec)connection. The DSLAM then routes ATM traffic to the home where the ATM traffic is received by a remote ADSL transceiver unit, while filtering off low-band plain-old telephone service (POTS).
The service user can request broadcast-quality video services by selecting a particular program, which is delivered using packetized compressed video (MPEG-2) with ATM as the transport mechanism. The beauty in this scenario is that MPEG compression technology not only serves to provision video over relatively low-bandwidth transport pipes, it is also a technology well suited to use ATM as a transport mechanism. ATM acts as a transport medium for the convergence of voice, video, and data, while DSL technology allows local access to ATM edge devices. ATM, MPEG-2, and DSL all act as complementary technologies to provide end-to-end multimedia services.
Even with the advances in domestic data-communication technology, it was difficult for anyone to imagine that broadcast-quality video services could be provided to TV sets over the same 24- and 26-gauge wires. However, the technology exists today to provision video services in the form of MPEG-2 video over this same infrastructure.
MPEG-2 technology has emerged as the de facto compression standard for distributed entertainment-quality video. It can efficiently compress full-motion video data for transmission over ATM networks. Full-motion digitized and uncompressed National Television Standards Committee (NTSC)-quality video requires a data transfer rate of about 240 Mbits/sec. With little perceived degradation, MPEG-2 can crunch this down to 4 or 5 Mbits/sec for distribution-quality video.
One of the greatest synergies between MPEG-2 encoded video and the ATM transport network lies in the fact that each of their respective bit structures is based on a fixed length. MPEG-2 packets comprise fixed 188-byte packets (184-byte payload plus 4-byte link header). That makes the logistics mapping of MPEG-2 transport over ATM simple. One convenient aspect of mapping compressed video onto ATM is that two 188-byte MPEG-2 packets, with 8 trailer bytes, exactly maps into eight 48-byte ATM payloads.
Video-aware edge devices serve as a bridge between video networks and packet networks, allowing telecom service providers to tap into the revenue opportunity of TV entertainment services. In the case of MPEG-2 video, the video-aware network edge device's ability to mitigate cell delay variation is of the utmost importance. ATM must cope with the needs of the data being transported and provision some function to manage each video stream's requirements. Getting MPEG-2 onto ATM networks, then picking it off in good order takes some care. The video-aware network edge device must be particularly adept at handling MPEG switching and jitter management to compensate for propagation delays in the network. Jitter management must include a combination of buffering, fixed-delay queuing, time stamping, and steady-rate outputting. Local MPEG-2 video streams are typically transported via an interface known as digital video broadcast asynchronous serial interface (DVB-ASI).
Although it was originally designed to transport data, Internet protocol (IP) packet-based technology can provide high-quality digital video integrated with interactive digital voice and data services over broadband networks. Trans porting multimedia such as digital TV or video-on-demand (VoD) over IP has many advantages.
It also has its weaknesses. For these advanced services, quality can become an issue in the IP environment. Emerging technology overcomes these quality challenges and enables IP to transport high-end digital video on one converged network. IP allows cable TV and telecom service providers to transport high-quality interactive digital video consumers want over their existing networks. It has become a leading contender for video transport, because the popularity of high-speed Internet access has spurred the creation of a broadband network that is reaching into homes everywhere. A connectionless, packet-switching protocol, IP provides packet routing, fragmentation, and reassembly, allowing many users of multiple services to share bandwidth. This new model of fast, multiservice delivery is a sharp contrast to the traditional community-access television (CATV) plant, which is a branch and tree topology that distributes content from a centralized headend out to the customer's premises. Traditional CATV headends act as satellite and microwave reception areas and drive a one-way TV broadcast infrastructure. The new model is bidirectional and therefore enables interactive services such as VoD.
Quality is always a priority when transporting video, especially for entertainment. The de facto compression standard for distributed entertainment-quality video like VoD and digital TV is MPEG-2 technology. The innate time-sensitive nature of video makes video stream clocking management every bit as important as content management. MPEG-2 injects into the stream a free-running 27-MHz timing clock called a program clock reference (PCR). MPEG-2 digital TV systems require that the encoder PCR time clock and the decoder's clock are kept in close synch. MPEG-2 compressed video is sensitive to propagation delay variance. Lost or corrupted data cause noticeable interruptions. That's why MPEG-2 has stringent quality-of-service (QoS) parameters.
IP is one option for the MPEG-2 transport protocol. However, IP was originally designed as a data transport protocol, and data is inherently more tolerant to error than media such as real-time voice or digital video. IP technology breaks a digital information stream into packets and gives each packet a destination. IP-based edge devices find a path through available network links. Packet delay variations, packet loss, and bit-error rate (BER) are secondary considerations. For high-priority data, where guaranteed delivery is necessary, protocols like transmission control protocol (TCP) determine whether packets have arrived successfully. If they haven't, the protocol initiates a retransmission. While that is fine for non-real-time applications such as e-mail, Web browsing, and file transfer, it inhibits real-time digital video. Digital video is sensitive to packet delay variation, packet loss, and BER. The high-bandwidth demands of video feeds make the problem worse.
Many point-to-point networks, such as those for phone service, don't have this quality-control problem. IP does not establish any comparable type of point-to-point connection. However, tagging different services to be handled differently by the network is a function of QoS that IP can accommodate. Services can be assigned classifications based on sensitivity and priority. IP can differentiate classes of service (CoSs) and distinguish different levels of service and network handling. That's the essence of QoS.
A number of differential traffic-handling approaches can provision IP-based CoSs, which in turn enables QoS. The DiffServ (differentiated services) standard allows networks to prioritize packets forwarded from the network edge devices based on defined service codes. Video, for example, would be coded differently than Web traffic. While that does not guarantee QoS, it does establish priority relative to other traffic on the network.
Because a typical IP network itself has no direct knowledge of how to optimize the path for a particular application or user, IP provides a limited facility, called type of service (ToS), for upper-layer IP protocols to convey what service tradeoffs should be made for the particular packet. The ToS field in the IP header is not considered true QoS, but it is still a way to identify special handling for a packet. Next-generation cable modems could use ToS to drive IP-based multimedia over their hybrid fiber/coax plants. A number of resource reservation protocols like resource reservation setup protocol and Multiprotocol Label Switching (MPLS) exist for next-generation systems to implement differen-tiated services, including video.
MPLS handles traffic management for media with different CoSs and injects predictability into the networks, dealing with QoS issues such as black-and-white management, latency, jitter, and packet/cell loss. MPLS gives IP the QoS and traffic engineering attributes it lacks; these attributes are needed to transport real-time video over burdened IP networks. MPLS in theory can provision QoS over multiple types of network technologies. The trick is getting MPLS provisioned end-to-end over multiple network types (ATM, IP, SONET, and DWDM). An initiative spun off of MPLS, called Multiprotocol Lambda Switching, would provision QoS on the fiber layer of the network. A standard regulating how MPLS should be provisioned end-to-end in the optical (SONET/SDH), ATM, and IP layers is expected to be finalized within the next two to four years.
Video-aware network edge devices in conjunction with MPEG-2 video compression, packet networks (ATM/IP/MPLS), deeper fiber penetration, and DSL technology will give telecom operators what they need to compete in the video delivery market: a single converged multiservice network that allows dynamic video, voice, and data provisioning (see Figure). The new telecommunications strategy is to evolve into a multiservice network over a broadband infrastructure ready for the addition of video entertainment services. As these services are rolled out, the consumer will reap the benefit of choice, content diversity, interactivity, and on-demand services, which will propel telecom providers to the top of the bundled service market because of their large UTP networks. Video is the key to this market breakthrough and will overshadow voice and data services due to its entertainment draw and revenue-generating ability.
Dave Pecorella is director of product marketing and Marty Dugan is director of corporate marketing at Artel Video Systems (Marlboro, MA). They can be reached at firstname.lastname@example.org and email@example.com, respectively.