Wireline carriers realize that in order to support IPTV services they can no longer incrementally enhance or "optimize" existing network elements. Unlike high-speed Internet access, IPTV must compete, in both quality and services breadth, with a long-entrenched incumbent standard bearer: cable TV. To compete successfully, carriers will have to overhaul their existing network architectures to enable greater capacity and flexibility. Unlike their cable competitors, whose network equipment investments suggest gradual optimization, carriers have a unique opportunity to leapfrog currently available offerings by delivering a radically improved TV viewing experience to subscribers. Features like Time-shift TV and massive Video on Demand (VOD) libraries can have significant impact on network design and are all the more reason why carriers must architect their networks wisely.

Bringing Home the Bandwidth

IPTV is a critical element of the digital home, an evolving term which describes the trend toward home subscriber environments, including inter-networked elements that enable seamless user experience regardless of content origin for TV, music, phone, and high-speed Internet. Enabling these new services requires the ability to deliver sufficient network capacity.
In order for carriers to predict bandwidth requirements and their consequences on equipment selection and deployment, it is useful to have a starting point for simultaneous service delivery to the digital home. A safe initial assumption would include two Standard Definition (SD) and one High Definition (HD) TV streams, three Voice over IP (VoIP) phones, and streaming digital audio/music. 

With advanced compression standards (H.264), this digital home scenario suggests a minimum bandwidth requirement of 15 megabytes . This is a conservative estimate based on technology available today, but this will surely increase as HD content becomes ubiquitous, and HD-capable displays become increasingly commoditized. This is critical for carriers to bear in mind because three simultaneous HD streams alone require 24 megabytes, without even considering the implications of upcoming applications, such as video telephony and personal broadcast. This could drive the bandwidth requirements for the digital home to 50Mbps and above, which in turn will have a number of implications for carriers on selecting appropriate access technology solutions.

Endless Channels, Multiple Delivery Options

Carriers have options for delivering this high bandwidth to the digital home. DSL technologies have grown at a tremendous rate in recent years and analyst projections suggest this trend will continue. Digital Subscriber Line Access Multiplexer (DSLAM) platforms must accommodate current bandwidth requirements while simultaneously providing a path for future growth. This means providing a backplane sufficient to support both today's and tomorrow's highest speed DSL variants. These include ADSL2+ (here today, with up to 26 megabytes downstream and 3-6 megabytes upstream), and VDSL2 (already in trials, with up to 100 megabytes downstream and 50-100 megabytes upstream). 

Next-generation DSL standards are an important consideration for carriers in their selection of appropriate DSLAM platforms. Equally worthy of consideration is DSL's inherent distance limitations and the potential for new technologies to provide remedies. Environmentally hardened DSLAM platforms designed for outdoor deployment are now available. These DSLAMs incorporate the latest DSL technologies in hermetically sealed, line-powered enclosures that enable carriers to effectively deploy DSL to subscribers and communities via 3kft and shorter copper loops, enabling data rates in excess of 50Mbps. 

Passive Optical Network (PON) technology provides a compelling complement to DSL for Fiber-to-the-Node (FTTN) applications, and/or a compelling alternative, in the case of Fiber-to-the-Home/Business (FTTH/FTTB) applications. An ideal solution where fiber is available in the access network, PON technology's symmetrical nature anticipates new types of interactive, bandwidth-intensive subscriber applications like video telephony. PONs combine the high capacity of fiber with the scalability of point-to-multipoint network topologies. Two PON variants stand to play an important role in access networks serving digital homes: Gigabit Ethernet PON (GEPON) is available today and ATM-based Gigabit PON (GPON) will be avail

able in the future. The currently available GEPON variant enables a single Gigabit Ethernet uplink to be split between 32 subscribers, affording 30 megabytes of symmetrical bandwidth to each connected digital home. 

IP Technology Frees Subscribers from Schedule-Driven Programming

Bandwidth capacity is not the only predictor of an operator's success in accommodating the requirements of the evolving digital home. The operator's access network platforms must enable flexibility and bandwidth management to an extent previously reserved only for equipment residing in the network core. Time-shift TV is an excellent example of a service that necessitates increasingly intelligent access solutions. Time-shift TV, also known as network Personal Video Recorder (nPVR), coupled with network-wide VOD, expands the concept of "watching TV" to include higher levels of interaction and control. Essentially, Time-shift TV enables subscribers to pause and rewind  interactive services like Time-shift TV. The downside of the unicast model is that the network must accommodate all of this new content and can become bogged down if it does not support multicast and unicast transition throughout (for both core and access). 

The dilemma for carriers is how to provide the greatest possible flexibility in their network so that whenever a subscriber requests a program that is already being delivered to other subscriber(s), that request will result in shared distribution of that program (the subscriber joins the appropriate multicast group). This will conserve bandwidth by eliminating the need for identical instances of a program to traverse the network. The Internet Group Management Protocol (IGMP) can solve this dilemma. 

IGMP protocol is an Internet protocol that enables DSLAMs, PON Optical Line Terminals (OLTs), and routers to passively "snoop" subscriber traffic in order to identify and properly assign multicast group membership. An access platform with this functionality checks IGMP packets passing through it, picks out the group registration information, and configures multicasting accordingly. Without IGMP snooping, multicast traffic is treated in the same manner as broadcast traffic, that is, it is forwarded to all ports. Via IGMP snooping, multicast group traffic is only forwarded to ports servicing members identified as belonging to that particular multicast group. IGMP snooping generates no additional network traffic, allowing carriers to significantly reduce network congestion.

Bandwidth management is another critical component of successful IPTV deployment (the most bandwidth-intensive deliverable to the digital home). To facilitate consistency with their network management and load-monitoring practices for other less bandwidth-intensive applications like telephony and high-speed Internet access, carriers must select IPTV systems with tools that enable them to carefully monitor and predict the results of network oversubscription rates. These tools must alert them ahead of time to adjust the oversubscription ratio/capacity to meet their service commitments. Average subscriber viewing behavior, relative to peak viewing behavior (major sporting events), along with unpredictable peaks and surges (disaster coverage), must be anticipated and bandwidth allocation made flexible. These monitoring and modeling tools must communicate with a comprehensive Unified Management System (UMS) that enables rapid alerts and response to potential network problems.

The Distributed IPTV Model

The carrier's goal is to minimize the amount of traffic that must traverse the core network. According to this logic, the worst possible scenario from a bandwidth management/utilization perspective for a large-scale deployment is a centralized model. Ironically, this model is often depicted in network topology diagrams illustrating IPTV deployments. In this model, a centralized "super-headend" combines encoders, back office servers, and VOD servers. Regardless of whether the network is delivering live TV, VOD, or Time-shift TV (essentially the same as VOD the instant it switches from multicast to unicast), all content and network traffic resulting from subscriber requests must traverse the entire network from the super-headend all the way to each subscriber's Set-Top Box (STB).

By changing their network topologies to encompass a flexible, distributed model, carriers will realize huge advantages, particularly for VOD and Time-shift TV. Unlike the centralized model described above, wherein single-source encoding takes place at the super-headend and is then multicast throughout the network, a distributed model utilizes regional headends so that local content (community interest / news) is only distributed in-region. This unburdens the core and the access networks. While an IPTV network does have the advantage of enabling local content to be viewable outside of region, there is substantially less demand for this programming outside of the respective region and would be more suitably delivered via unicast stream.

Most of the actual traffic that traverses an IPTV network is going to be from live TV and subscriber requests (authentication / billing information / Electronic Programming Guide (EPG)). Like encoding, authentication, EPG, and the billing can be distributed as well. 

A final component of successful IPTV delivery concerns the content storage and distribution mechanism itself. Since maximum distribution is the key to creating flexibility, it stands to reason that a segmentation scheme, in which sequential content segments are distributed piecemeal, would radically reduce network congestion. Rather than sending multiple iterations of content in their entirety, from one storage and streaming server to another in a segmented content solution, adjustable segments replace large files. Content segmentation at the edge, coupled with innovative protocols like Broadband Media Distribution Protocol (BMDP), which enable the IPTV system to adjust storage and distribution heuristically according to trends in subscriber behavior, provide a buffer against network jitter and greater tolerance for peak bursts in traffic.

IPTV means new revenue for carriers, and its differentiated services suggest a huge potential benefit for both carriers and the subscribers they serve. With the coming of virtually infinite VOD and Time-shift TV, the industry is poised to witness the first radical advances in television since digital cable. However, while subscribers are accustomed to occasional service quality variance due to traffic latency for streaming video applications on their computers, they will have limited tolerance for similar issues that affect their long-familiar TV watching experience. After all, people do not subscribe to services based on technology, they subscribe based on quality of service, value, differentiation, and convenience. IPTV requires careful selection of new technology solutions that will ensure successful initial implementation, and scalability, enabling carriers who plan correctly to move on to their own cycle of optimization.