 |
Intelligent Content Distribution: A Critical Enabler for IPTV
September 1, 2006
Everybody is talking about IPTV these days and for good reason. Video is the stickiest service of the triple play, and service providers are looking desperately for ways to increase revenue while fighting to retain customers.
Cable operators have been quicker to market with triple play packages, largely because voice is easier for them to add than video is for a telco. To date, satellite has been the telcos’ video route to market, but with the advent of IP and multimedia services, telcos cannot be competitive by simply maintaining the status quo.
As cable-based triple play offers continue to steal away telco subscribers, there’s an increasing urgency for telcos to roll out IPTV. Not only are telcos at risk of losing their traditional revenue streams but also the new revenues that would come from video services.
Besides aggressive cable operator marketing efforts, broadband adoption is growing rapidly in the United States, which is a prerequisite for IP-based services. In addition, mobility is becoming ever-present, and the market for mobile video services is beginning to emerge. Never before have consumers had so many video options; the battle for the broadband home may become the most important challenge telcos have ever faced.
Haste Makes Waste
In their haste to implement IPTV, telecom operators may risk compromising end-user functionality by architecting a point-to-point system. Currently, many IPTV deployments have a video server with limited capacity at the network edge communicating directly with the set-top box in the customer’s home. This silo solution eventually will need to be connected to the network core as IMS solutions are implemented.
In so doing, it will be critical to optimize distribution of content throughout various network parts to minimize congestion and maximize bandwidth. Carriers will need to move content dynamically to various storage and video servers throughout the network and cache them according to demand.
As IPTV deployments expand and network capacity needs to scale, this approach can become problematic. When telcos expand into new markets or neighborhoods, they need to replicate their models by adding more video servers, which in turn delivers content to the subscriber’s set-top box. This is a costly, inefficient approach, especially for utilizing content storage.
Subscribers suffer as well: Edge servers have limited capacity and cannot directly handle all viewing requests. Less popular choices take longer to access; latency will be noticed and unwelcome.
Much like the five 9s quality of the PSTN, IPTV subscribers expect to receive broadcast quality service from the provider. This is one area where telcos cannot come up short.
The Internet was designed initially for sporadic bursts of data, which it handles well. As the world moves to IP and new applications are introduced, network bandwidth and managing network bandwidth become more of an issue, especially for real-time communications.
Although the rate of network capacity innovation is high and technologies such as Gigbit Ethernet and 10GigE are evolving to help address this issue, bandwidth intensive applications like video and specifically IPTV pose a challenge.
Delivering both streaming as well as on-demand content from the network core no longer uses sporadic bursts of data; rather a constant high utilization connection is required. With several television sets per household, this will tax severely both the access infrastructure and the core network.
Intelligent Distribution
Intelligent content distribution is the key to alleviating these network choke points. This approach will also expedite delivery content to the user and minimize latency, which is becoming increasingly important as content increases in size, scope and numbers.
In 2005, 584 full-length motion pictures were released compared to 263 just 10 years earlier. This rate of new movie releases is expected to continue as more smaller-budget films are introduced, and capital continues to flow into Hollywood. This not only applies to IPTV but to all digital content.
According to IFPI, online song catalogues doubled in 2005 to 2 million tracks, available on 335 legal download sites (up from 50 two years ago). Sales of music via the Internet and mobile phones grew more than 289 percent in 2005, generating more than $1.1 billion in revenues.
With the move toward digital music, the long tail effect has the potential to increase dramatically the overall market size by easily exposing new artists. In fact, the move toward digital music has already begun. According to the RIAA, digital units sales increased 166 percent from 2004 to 2005, while the sale of physical units declined 8 percent.
Comparable data for voice is sporadic, but we believe the impact of digitization and IP on the video and music industries is highly indicative of what is coming. It is not a great leap to conclude that peer-to peer applications such as Skype are increasing the pie for voice minutes, rather than cannibalizing TDM traffic.
Similarly, with the proliferation of SIP and IP-enabled endpoints (e.g., PCs, mobile phones, PDAs), demand invariably will increase for a growing variety of video content. Some content will be in existing forms, but several new forms will emerge that can only exist in a broadband world. Clearly, we are entering a market driven by economics of abundance rather than scarcity.
This move to digital presents a great opportunity for service providers to act as a central repository for all digital content: voice, data and video. As part of their IMS strategy, this centrally stored digital content can be consumed on any device over any network.
Although admittedly there remain some DRM concerns, consumers want the ability to download songs on their phones and listen to them later through their PCs. Similarly, consumers want to watch their favorite shows on their HD TVs, laptops, or on mobile phones.
Too Many Formats
Without real-time codec translation, however, each piece of content will have to be translated and stored in many formats. For music, this means at least four different formats (e.g., Windows Media Player, Real Player, iTunes, MP3) and also for video (e.g., common interchange, quarter-common interchange, subquarter-common interchange, and advanced video coding).
This requires a great deal of storage, which typically leads to non-utilized servers and management issues, while delivery of content can lead to network congestion.
These issues can be minimized by properly designing the IPTV architecture and by distributing content storage throughout the network. At the root is the reality that not all forms of content are created equal. The demand and viewing habits for live programming are very different from prerecorded or archived content.
Similarly, the storage requirements for a hit TV show will be very different from that for an obscure show or movie with a tiny following. In this decentralized model are four distinct nodes or hubs for storing various forms of content:
- Home — Local DVR: Stores the most recent shows and favorites.
- Home — Media Server: Stores all purchased and downloaded content (e.g., movies, TV shows, music).
- Network Edge DVR: Stores the most popular content (e.g., most downloaded movies, new releases, most popular music).
- Network Core DVR: Stores all content in all formats (e.g., all movies and music available for download).
Another Complex Layer
The end-user experience adds another layer of complexity to how storage is best addressed for IPTV. Until recently, video has been watched exclusively on the television set. Viewing options and formats have increased considerably, even just for TV, which now includes rear projection, plasma, home theatres and most recently HD.
Today, the PC rapidly is becoming a bona fide alternative, both in the home, and, thanks to technologies such as place shifting, on the road as well. The inherent flexibility of IP means that the viewing experience can be highly personalized; much like how cable networks created more programming choice, IPTV will take this to a whole new level. IPTV is tailor-made to support personalized content, but a proper storage infrastructure must be in place.
Within U.S. homes, the growth of DVRs has begun with more than 2 million units shipped in 2003. IDC expects this to grow to 18 million by 2009. By dynamically moving content around the network and caching it in various locations, operators can optimize bandwidth utilization, reduce the choke point, and deliver a better user experience with little latency. In addition, it will present a major cost-saving opportunity.
By having the ability to move content around dynamically, operators can transfer files that have become stale to less expensive storage devices, while keeping the high-end storage facilities for high-value content.
Swarmcasting
This can increase the number of lower-end storage devices while ensuring that utilization rates are close to capacity. Furthermore, it is possible to improve this scenario dramatically by introducing swarmcasting technology for content caching.
Under this scenario, the video content can be broken up into 30-second chunks and cached in set-top boxes around the neighborhood. When a particular movie is requested, the policy enforcement point aggregates all the chunks and reassembles the original content piece to deliver to the user.
In addition, focused content may be cached in areas where emphasis is on a personalized user experience. Storing advertisements on home DVRs based on user behavior, for example, makes a lot of sense since this content is geared specifically to a certain audience.
This scenario can be applied to either on-demand or streaming content. In either case, the large size of video files dictates a more efficient delivery mechanism, especially given that many broadband services offered by telcos is DSL, not fiber.
In the U.S., many service providers are pursuing fiber buildouts to support their IPTV strategies. Verizon is pursuing an end-to-end FTTH (fiber-to-the-home) deployment with their FiOS strategy, while AT&T’s Lightspeed offering has opted for a FTTN (fiber-to-the node/ neighborhood) strategy.
In the interim, many service providers are pursuing IPTV deployments that will run over DSL networks. This will be problematic for telcos when households have more than one HD TV set. DSL cannot yet deliver enough bandwidth to support personalized content delivery at this level, and in these cases, IPTV will have some limitations.
IPTV implementations are only in their infant stages, but the issue of intelligent content distribution needs to be addressed now. The true value proposition for triple play does not lie in the convenience of bundling three distinct services as a one-stop shop.
Rather, the value is in the integration of these services on a common plane, and more importantly, the ability for subscribers to create their own bundles as well as customize and control the end-user experience.
Given the value video brings to the bundle, storage and content distribution are vital enablers to make this possible, and the sooner telcos address distribution, the sooner they will be able to keep pace with the competition.
Danny Klein is an associate at Vesbridge Partners LLC. (dklein@vesbridge.com)
Jon Arnold is the principal of J.Arnold & Associates. (jon@jarnoldassociates.com)
Related Member:
Dan Klein
|
 |
 |
|