June 18, 2002

The Future of Video: The Implications of Hyper-Cheap Storage

Stewart Alsop is always worth listening to. Here he serves up his vision of what massively plunging storage costs mean for the future of video entertainment. It's an interesting race--between computation, storage capacity, and bandwidth. I'm becoming convinced that what kind of future we live in depends on which becomes cheapest fastest.


ALSOP ON INFOTECH: I Want My File-Served TV!

Let me offer up a vision of the future of television. I'm a director of TiVo, the publicly held company that makes those personal video recorders (PVRs) that allow you to record TV programs onto a hard drive to watch whenever you like. (My venture capital firm, NEA, was an early investor in TiVo but no longer owns shares of the company. I personally own shares.) At last week's board meeting, another director, John Hendricks, the chairman and CEO of Discovery Communications, brought up the phrase "file-served television." I first heard Hendricks say those words two months ago, and they've been bugging me ever since. Last week I figured out why--it's because file-served television seems the only sensible future for television. If I'm right, that's good news for TiVo, and for consumers as well.

The vision du jour of the future of television is video-on-demand (VOD)--broadcasters or distributors of television have all video known to man stored in digital form, so any viewer anywhere can instantly watch anything at any time on his TV. As we speak, several big cable companies are spending oodles of money installing VOD systems. One recent estimate is that VOD is available for about seven million households in the U.S.

Now, let's stop to consider that the U.S., every day, produces something on the order of 360 hours (12 hours times 30 channels) of new content (both theatrical movies and television broadcasts). That's about 130,000 hours a year. Take into account global output, throw in historical footage, and you're quickly up to a million hours of video programming.

The biggest video-on-demand system being deployed by cable companies this year will have 1,500 hours of video stored on its centralized computers. A lot of smart people think scaling beyond that will be terribly difficult, which means this kind of VOD isn't much of an advance. It will serve a narrow funnel of content to consumers, just as content programmers now decide what viewers get to see.

So if you believe that people do want to watch whatever they want whenever they want, you need a massively distributed system. Which brings us to "file-served television" and that TiVo board meeting. Having a PVR's really big hard disk in many living rooms creates a massively distributed system: Instead of relatively few hard disks owned by the cable operators, you have hundreds of thousands of hard disks owned by everybody. And thus the space to store a million hours of video content.

The problem is that movie companies and television companies aren't thrilled to have their valuable stuff sitting on hard disks in your living room. They think they might lose control over how their intellectual property is sold and how they make money. John Hendricks knows all about this, given that Discovery Communications depends on people paying a premium for content like When Dinosaurs Roamed America. He doesn't have an answer to the payment problem, but he's confident enough that one will emerge to propose this seemingly risky plan.

Essentially, file-served television describes an Internet for video content. Anyone--from movie company to homeowner--could store video on his own hard disk and make it available for a price. Movie and television companies would have tons of hard disks with huge capacities, since they can afford to store everything they produce. Cable operators and satellite companies might have some hard disks to store the most popular content, since they can charge a premium for such stuff. And homeowners might have hard disks (possibly in the form of PVRs) that can be used as temporary storage for content that takes time to get or that they only want to rent--or permanent storage for what they've bought.

Like caching on the Internet, this system is designed to push the more popular content closer to the user. So if you want to watch reruns of Gilligan's Island, you might order them from the one copy maintained by Turner Broadcasting Systems' hard disks in Atlanta and have it delivered through the system to your TV in Boise. But if you want to watch the new Lord of the Rings next December, you'll need to reserve a copy from Comcast, which will download it to your living-room hard disk the night it's released with a key that lets you watch it a certain number of times.

That's file-served television. It is very different from today's TV: The popularity of content is controlled by users rather than broadcasters. It's a system flexible enough to adapt to a new TV standard like HDTV over time. It's a vision that's big enough, in fact, to contain all the previous visions of the television industry.

Posted by DeLong at June 18, 2002 10:21 PM

Post a comment