BUILDING THE INTERACTIVE TV INFRASTRUCTURE

There was a time when you could measure the size, if notsophistication, of your hometown by how many TV stations you received.Today, that’s all changed. Before those of us raised on channels4, 5, and 7 retire, there will be more TV choices than there areviewers.

Accept that premise and you’ll see a communications industryturned on its head. Channels lose their meaning when viewers can selectprograms directly. Programs could be produced for interested viewers,not just to appease the lowest common denominator. Imagine terabytes ofbandwidth filled with interesting programs. Before you dismiss this asperhaps another George Gilder wave division, multiplexed pipedream,consider that you can already see evidence of the development of thisinfrastructure.

By now, you’re probably thinking I am a new media optimist,ready to lavish praise on the TV killer app du jour. So let meshift gears with a quote:

“Television is the supreme time waster of all time,”said Jonathan Tapin of Intertainer, a video on demand service, at lastyear’s Digital Hollywood convention. Of all the exciting newbells and whistles promised by the interactive television hype, Tapinsaid, “First, give me what I want, when I want it.”

What I extract from Tapin is support for a long-held belief. Whatpeople primarily want from their television is television programming:that’s full-screen, full-motion video images with sound thatmatches the lips. It’s striking how many millions of dollars havebeen invested in companies that provide some TV function other thanthat! People may want their program to start when they want. They maywant to pause it or skip ahead when they want. They may even want thatprogram to be compiled in a manner that is uniquely relevant to them.But it is still a TV show.

Just to avoid a lot of bothersome apologizing, I should admit thatpersonalizing the stock ticker on my TV screen, buying a pizza with myremote control, or playing along with Jeopardy! are all swellthings that some genuinely clever people have figured out how to do.These too will be part of the future. But there’s a reason theycall such things enhanced television. The main lure isn’t theenhancement; it’s the object being enhanced—namely, thefull-motion, full-screen pictures and sound. In this grand future,where will these come from? Who will own them? How will I select them?How will the money be made?

These answers are more straightforward than you might guess. For atleast the near future, the primary television programming people aregoing to watch are the same programs they’ve been trained towatch for all these years. Sure, videochat may one day reign withBaywatch 2010 atop the Nielsen ratings. But while videochat andits illicit cousin, VoyeurCam, will become the ultimate (full-screen,full-motion) reality programming, that doesn’t mean our appetitefor all the other staples (movies, sports, news, sitcoms, soaps, etc.)are doomed to become artifacts of a lost culture.

A Brave New World

To figure out how these programs will get from the corporations whoown them to you, just subtract those parts of the current deliveryinfrastructure that have outlived their usefulness. If I’mParamount, I traditionally had but one avenue to make money from my 168episodes of Hogan’s Heroes. I licensed a TV station to runthe series within certain guidelines, and then the station paid me. Thestation sold commercials, paid its expenses, and pocketed the profit.This model existed because there was never an alternative. Morerecently, distributors have had the option of skipping the middleman.Thus, Hogan’s Heroes is on TV Land, a cable channel ownedby Paramount’s parent, Viacom.

The next step is to augment, though not necessarily replace the“packaging” function performed by the TV station. Themoment I’m ready to watch Hogan’s Heroes, I can plopdown on my couch, and navigate to a menu that lets me chose the veryepisode. It might even be a TV Land branded menu. I click, it starts.Full-screen, full-motion video. I would argue that many viewers wouldbe willing to pay a nominal charge for this service. If not, manyadvertisers would be willing to buy the first few spot positions in mybroadcast day, whenever it may start and whatever programming it maycontain. Direct mail meets television.

The first steps along these lines are happening now. In a handful ofcommunities across the United States, cable operators are introducingvideo-on-demand (VOD) movies. Many cable systems offer pay-per-viewfilms today, but those are scheduled services. Pay a couple of bucks,and you can select among a handful of titles that start at certaintimes. These new on-demand services, as the name suggests, enable theuser to select what movie he or she wants to watch, then play it at thedesired moment, and even pause it when nature or telemarketerscall.

On-demand movies are one thing. Watch next as the A&E cablechannel, the latest in a small handful of cable channels dabbling inthis area, starts placing some of its original programming on cable VODservers. Many years may span from when A&E accepts its first VODpurchase of a Biography rerun, to the day NBC premiers the latestepisode of Friends (or whatever that era’s Friendsmight be) on a menu-driven service. These years will mark a transitionperiod from television viewing that is fundamentally based onbroadcasting to viewing that is fundamentally offered on-demand.

The Underlying Infrastructure

This change alters the infrastructure that’s supposed todeliver the programming. Although we’ve got a lot to learn aboutusable design for television, our experience with the web suggestsit’s not a technological stretch to assume TV viewers mightnavigate among menus of content companies, drilling down first to aspecific series and perhaps even to an individual episode. The lessintuitive part is trying to understand what system will bring thiscontent from wherever it is stored to my local headend or centraloffice, and then on to my TV in glorious full-motion, full-screenquality.

Let’s tackle the wholesale problem first: how the content getsto a vendor your TV can talk to. Currently, if you’re among thefortunate few with access to video on demand from your cable system,you can choose from the titles that are physically stored in your localcable head end. As a result, you’re getting the equivalent ofBlockbuster’s Current Releases shelf. (Well, almost current:cable VOD windows are a whole separate story.) That makes sense:there’s a limited number of homes accessing the material,there’s a limited number of disks to store material upon, andthere’s a limited upside to parking less popular items on aserver with restricted reach.

Logic and business sense tells you there will always be finitelimits on how much content can be stored in the local server. Yet italso tells you that spread across a national audience, virtually anyprogram or film ever made has, even at 50 cents per play, the potentialto make more money than whatever it costs to park it on a server andoffer it for sale.

Here’s where we revisit a familiar graph: the quality of videocompression is improving; the cost of bandwidth is decreasing.Companies like Akamai, I-Beam, Enron and many others have begun tobuild backbone networks capable of connecting server farms operated bycontent owners with servers located in cable head ends or telephonecompany central offices. Caching concepts are so well understood thatit takes little imagination to contemplate a network where the mostpopular content is stored as close to the users as possible, whileperhaps just the first few minutes of other material is close at hand.The system can go fetch the rest when it knows a user wants it.

But that still leaves the famous “last mile” problem.Internet-style streaming media isn’t good enough. Few of us haveenough Internet bandwidth to receive enough bits fast enough torepresent full-motion, full-screen video. So you need a fat pipe.Furthermore, you need a fat pipe that connects to the television.Lecture me all you like about the Internet generations. In my opinion,the device for watching entertainment programming in the home is thebox with the large screen that sits in front of the couch or on thekitchen counter. People change more slowly than technology.

So what we need is a delivery infrastructure that enables each andevery television viewer to potentially watch something different, atthe same time. Not 500 channels, but one channel per television set. Atthe moment, four different kinds of companies are competing to connectyour home to these types of services.

If you’re a telephone company, or a cable overbuilder, the onechannel per TV paradigm is not a future problem, it’stoday’s problem. Using XDSL and Ethernet, respectively, thesevendors’ pipes are much thinner than cable TV or direct broadcastsatellite (DBS). Thus, instead of sending your TV or set top box allthe channels and letting you tune to the one you want, the XDSL andEthernet vendors will just send you the channel you select, in effectoffering a remote control channel change where your clicker actuallyswitches a tuner located across the street or even across town.

In this model, even ESPN Sunday Baseball is an on-demand program:you don’t control the start time, and you may not pay for itdirectly, but the program isn’t delivered to you until you askfor it. That means these are the guys most ready for the future:they’re already delivering one program for every TV set. Butwidespread rollout of video over DSL or Ethernet is stymied by a forcemore powerful than technology -- money. Construction of both newsystems and press releases touting new systems have slowed dramaticallysince the NASDAQ’s decline.

Direct broadcast satellite operators have a lot of bandwidth, butnot enough to send each subscriber his or her own channel. Furthermore,there’s the troublesome problem of the return path. To evenconsider responding to a viewer’s click, you first have to beable to receive that click. Using a telephone line works for nightlybilling queries, but not for real-time operations like pausing anon-demand movie.

New tools are on the horizon for DBS, but each has its own quirks.Ka-Band satellites offer “spot beams” so that individualchannels could be re-used dozens of times across the United States.Still, that’s unlikely to provide enough bandwidth to support theone-program-per-TV paradigm. Two-way satellite communications areavailable, but the dishes are slightly larger and would requireconsumers to swap their existing equipment. Personal video recorders(PVR) are the Great White Hope for DBS video on demand: store themovies in the consumers’ homes, but only “unlock”them when purchased. Already both major US DBS vendors are offeringreceiver boxes with PVRs inside. The marketers’ thinking is thatmaybe combining these functions within one box will finally openpeople’s wallets. Thus far, PVRs clearly represent the mostlife-changing home entertainment technology that few are willing tobuy.

Your cable television operator, on paper, has the return path issuesolved and enjoys bandwidth that can be expanded dramatically. Inpractice, many systems still have a fair amount of money to spend toupgrade noisy systems and rollout digital services to the bulk of theircustomer base. Cable’s hybrid fiber coax (HFC) architecture usesfibre optic cable to send signals from the cable headend to“nodes’ located in neighborhoods. These nodes—reallyjust pole or street-mounted boxes—take the signal from the fibreand send it down copper wire (coax) to subscribers’ homes. Eachcoax cable might feed 500-5000 homes. While the copper coaxial cable toyour home limits the cable system to a finite number of channelpositions, statistical multiplexing and modulation techniques can beused to fit 10-16 digital programs in each channel slot.

Today’s video-on-demand systems are only beginning to exploita compelling advantage of the HFC infrastructure: a differentassortment of channels can be fed down the fibre that feeds eachneighborhood node. Thus, channels can be re-used from neighborhood toneighborhood without interference. When you order your pay-per-viewmovie from the comfort of your armchair in West Central Anytown, thecable system finds the movie file on its server, and then finds achannel slot that is not full that is feeding the West Central node.For example, the system may designate your movie as the tenth streaminside West Central channel 97. It then tells your set top box to tuneto that particular stream, and enables your set top box (but no others)to disable the encryption of that stream so that the scrambled filmbecomes unscrambled. All of that happens today, but only in the handfulof cable systems offering video on demand “trials.”

Slow Going

Truth is, the slow pace of change is worthy of note here. Not onlyhave most U.S. overbuilders and XDSL vendors abandoned or delayed videoplans, cable operators--the only players left standing who have enoughbandwidth to deliver true video on demand--are on the cusp of making atragic discovery: in tight economic times, VOD is not a very goodinvestment.

There are a number of reasons for this. The costs of provisioningtheir networks for all of those “personal channels”can’t be offset by the income generated from the limited array ofrecent but not-quite-new film titles the movie studios have beenwilling to license. Personal Biography reruns are novel, butdon’t radically change the return on investment on a $15 millionsystem. The “interactive television” applications that weresupposed to help justify the infrastructure have been slow tomaterialize. And those far-reaching wholesale content deliverynetworks, if successful, could essentially make today’s VODequipment investment obsolete.

For viewers, what all of this suggests in the near term is more ofthe same. Everybody knows there’s a big bold future out there,but no network operators want to write the big check. Change will comenot in forklift plant upgrades but in incremental equipment purchases.One box links the home delivery network to a wholesale content networkbrimming with video programming. Another offers content from a clevermovie studio. A third integrates data or voice services. Networkoperators will thus be empowered to spend their money on individualapplications that they think can make money, instead of painfulinvestment in “infrastructure,” for which stockholders nowdemonstrate limited patience.

When those “boxes” are integrated with common managementsystems, the future begins. And the television industry will changedramatically, for the better.

Chris Lee is Director of Marketing, Advanced Platforms forTerayon Communications Systems. He can be reached atclee@terayon.com.