Packet technology for ìTelco TVî Service Providers

Given its thirst for bandwidth, video is viewed as a potential driver for higher capacity multiservice networks – even to the extent of once being hyped as the next ìkiller applicationî.

In reality, video, as part of the so called ìtriple playî, delivered over packet technology has been slow to make its mark on both the residential and SoHo markets.

There are a number of reasons why this has been the case in the past but the important thing is that the sentiment is changing.†Telecom service providers have an opportunity to use the existing telephone line as a broadband entry point to the residential market and SoHo market. The continued rollout of ADSL is creating, each day, a larger target market for much more than just a high speed (and flat rate) Internet pipe.

Studies in the CAPEX costs of greenfield residential video deployment shows that the cost per subscriber for the ìtelco TVî model can be in the range of $1,300-$1,500 ñ compared with a figure approaching $4,000 for cable delivery[1].†

Even higher speeds can be delivered by Fiber to the Home or Premises (FTTH/FTTP), which is now being rolled out in pilot trials.†

The motivation for the service provider can be seen by examining the three business benefits:

  • New application, new customers : For these service providers, video is a potentially new revenue stream to add to existing voice and data revenues. This means new customers and additional revenue from existing customers.

  • Reduced churn : Better still, the creation of additional services allows for service bundles and both the cable operators and telcos alike have found that bundling of services not only increases the revenue per household but also reduces the churn rate. Even bundling two services reduced churn rates by 33% as compared when just a single service offering had been taken. [Based on reports by ECI Telecom and Minerva Networks]

  • Competitive edge : Lastly, by providing video services, the telco is directly positioned to compete with the cable operators that have traditionally been the sole provider of video services.

    In the residential market, a telco looking to deliver video has a much simpler proposition to package and sell as the target audience is already typically a subscriber to a mix of basic and premium TV and video services.
    However, itís a fair question that if the user demand for video remains as strong as ever, then why hasnít the ìtelco TVî model become more commonplace.†

    This is where technology does come into play and relies, quite simply, on four technology related enablers to come together.†

  • Broadband network technology delivers increased bandwidth per user over the existing infrastructure

  • Multicasting technology allows a single video stream to be fed to thousands of subscribers

  • Video compression technology delivers higher quality for a given amount of bandwidth

  • Price/performance improvements deliver affordable technology solutions to reduce the cost per video stream handled

    As an industry forum, that has been working on ATM specifications for building service provider multiservice networks for the last decade, one of the areas of interest is how these packet networks can realistically deliver video services, primarily to residential and SoHo business users.

    As a technology forum, the easiest mistake to make here is to see this purely as a technology issue. True, technology acts a major enabler and we will look at some ways in which technology in this space is evolving. But one canít overlook the issues of user demand and a viable service provider business model.†

    For The ATM Forum, transporting video over an ATM network is actually quite well established — a 1997 ATM Forum specification standardized the carrying of an MPEG-2 video stream over ATM. Recent developments in MPEG-4 video streaming and the H.264 specification have made the case for video over packet even more compelling.[2] Even without these new encode and decode techniques the typical trend, over time, is to reduce the bandwidth required while maintaining perceived quality.

    At the same time the available bandwidth is increasing anyway. Contributions are now being gathered for the transport of an MPEG-4/H.264 transport over ATM, following on from the earlier work to support MPEG-2 streams.†

    The telco TV model is based on a switched video model compared with cableís broadcast model. The switching of video is necessary as, even with the best compression today, only a few channels could be transported down the ADSL line concurrently – certainly not the 200 channels that a cable TV operator offers.

    In technology terms this requires that a subscriber can select and switch in real time to a different channel and for this the technical solution is multicasting. Here the user switching channels is actually initiating a request to join a different multicast group.

    The remote control is not just instructing the set top box but actually forcing a change back in the network.†

    Although most video streams will be agnostic towards the underlying transport (using their support for IP as a unifier), this does not imply that the network and application designer can ignore the specific needs of video as an application.

    Video applications need the network to provide guarantees of QoS including predictable latency and jitter. It is here that packet technologies like ATM are best suited to provide the infrastructure for this support.†

    This is not a new concept. There have been a number of previous VoD trials dating from the mid 90s. ñ many having broken new ground at the time. One of the first VoD trials in the world to go live with ATM protocols carried all the way to the home was in Cambridge, England. The architecture was fibre-to-the-kerb, with ATM switches installed in roadside cabinets.

    Given the lack of broadband DSL infrastructure they mostly used alternative access/delivery networks. [In that trial the final drop was actually over CATV coax cable, multiplexed with the TV signal].

    Each ATM switch supplied six households with a 2 Mbps bi-directional feed, carrying MPEG-over-ATM for the video, but also capable of more general traffic such as IP, for example for Internet access.

    Time Warner experimented with VoD as part of their Full Service Network (FSN) trial in 1997.†

    More recently, developments in fiber deployment have extended the case for video over packet beyond that of just the ADSL deployment. Fiber to the Premises (FTTP) is now gaining ground, overcoming some of the concerns of high deployment cost by the use of Passive Optical Networks (PON).

    BellSouth, SBC Communications and Verizon Communications have formed a Joint Procurement Consortium (JPC) to shortlist potential vendors following their joint RFP issued for PON based FTTP infrastructure earlier in the year. Indeed there are already standards in place with ITU-G.983 defining a standard for ATM based PON allowing delivery speeds of up to 622Mbps.

    This opens up huge potential for video applications with bandwidth available far exceeding that of any copper based local loop access (and perhaps more importantly that of the cable TV infrastructure).†

    So, why is success now more assured? The reason why the early pilots did not evolve into commercial offerings was that they typically lacked some of the key technology enablers and suffered from a less than compelling enough business case.

    Often, the technology to deliver video was simply too expensive to be applied to mass market applications like Video on Demand or Interactive TV. This meant that where video was being delivered over a packet infrastructure, it was limited to high value business applications like remote surveillance, distance learning and video conferencing.

    Things are different today, not just in technology terms, but because the customer is simply more savvy towards VoD, Tivo ô style PVR and interactive TV.

    So, in the world of video delivery, technology and business drivers are coming together to create a viable business model, especially for the local loop providers.

    Video delivery over packet networks today is a far cry from the early attempts to do low bandwidth video conferencing and music video clips over the Internet. Old misconceptions are being replaced with todayís reality ñ packet video is coming to a screen near you soon!†

    This article written by Marlis Humphrey references the ATM Forum Broadband Exchange.
    This is a series of events each addressing a topical industry issue. For more information then visit the ATM Forumís web page for the Packet Video Broadband Exchange in order to view the individual presentations from the event.† ATM Forum