Coming eventually: TV on the PC

When video over the Internet first made its debut some four years ago, it was slow, jerky, grainy and small. For most Internet users today, it's still slow, jerky, grainy and small. But despite the promise that broadband-essentially high-speed Internet connections-would bring TV-quality video online to the masses, the waiting continues.

Predictions vary as to when the world will have a true television experience on the PC, and opinions differ on the magical bit rate that will make that happen. But one thing is certain: There's a race to get there first.

In the near term, the competition between the two dominant providers of broadband service promises to be interesting. The cable industry has a massive lead on the phone companies' digital-subscriber line (DSL) platform. But then, cable had a head start. Of all U.S. households with high-speed Internet connections, 75% are signed up with cable, while 25% are wired for DSL, according to Cambridge, Mass.-based Forrester Research.

But DSL is gaining fast and is widely expected to catch up with or even surpass cable by 2004, according to Forrester. Look for wireless residential broadband including satellite, MMDS and fixed-wireless local loop in the arena as well, but those platforms are expected to serve only niche markets in geographical areas not served by cable or DSL.

"There's an ongoing fight about who is better," says Mark Zohar, research director for Forrester. "But it comes down to this: Nobody cares. People just want broadband. Whoever gets to the market first wins. If the cable company is there with cable-modem service, people will get that. If there's DSL, they'll get that. If there's both, then there's competition, and it will be whoever has the best price, service and brand name."

Indeed, the demand is there. According to a Forrester survey, 55% of traditional dial-up users say they are prepared to switch to broadband service once it becomes available. But with only 3% of the nation's 100 million households wired for high-speed connections, the waiting game will continue for most Americans, who are rapidly going online.

More than half of U.S. households, nearly 54 million, now have computers, and about 42% of those have Internet access, up from 26% in 1999 and only 19% in 1998, according to a Commerce Department study of computer usage released in October. The FCC's most recent report on broadband access, released in August, refers to predictions that the number of households online will double within five years, encompassing two-thirds of all U.S. households.

The reality is that cable and DSL are plagued by problems, including network constraints that keep the services out of reach for most consumers. The big issue: the so-called "last mile" to the home.

Although cable networks today employ a combination of new fiber-optics and traditional coaxial cables-a hybrid fiber coax-the same coaxial wires that have been in use for years still complete the connection to the home. Upgraded electronics, however, do help boost capacity.

Cable's main drawback

But cable's big weakness lies in its party-line nature: The greater the number of customers logged on in a neighborhood, the lower the bandwidth and the slower the stream. As MSOs sign up more customers for modem service, the shared-network design is a growing problem. But the industry has come up with a solution to the issue: node splitting. Fiber from the network terminates in nodes located on neighborhood street corners where coax takes over and runs into the home. Fiber-optic lines can be split at the node, creating a second line to relieve the original fiber of half its load.

With DSL, users effectively have their own dedicated service directly from their home to the local telephone company's network. As a result, DSL service does not deteriorate as more users sign on. But video service can be pushed only over short distances, no more than 12,000 feet of line from the central office. SBC, however, is addressing that problem, among others, as it spends $6 billion over the next three years to rebuild its plant for the digital age. By laying fiber to what it calls "neighborhood broadband gateways," or remote terminals, and copper to the customer's curb, the company is extending its reach to homes that otherwise would have been outside the service area.

Another difficulty telcos face is that copper lines upgraded with digital-loop carriers (DLCs) do not tolerate the massive compression needs of DSL and must be replaced with fiber. Bell South, in particular, must deal with this issue, since up to 30% of its network has been built out with DLC.

Satellite did not offer true broadband connections until early November, when StarBand Communications-a joint venture of Gilat Satellite Networks, EchoStar Communications and Microsoft-launched a two-way, always-on, high-speed satellite service for consumers. Hughes Network Systems will be next in early 2001, when it upgrades its one-way system-DirecPC-to two-way.

Late next year, iSky-a fledgling satellite company owned by Liberty Media, TV Guide, TRW, EchoStar and TeleSat-is also scheduled to deliver two-way high-speed data to residential markets. That venture intends to use Ka-band spectrum and spot-beam technology. And WildBlue, a Denver start-up, plans to get into the two-way game in 2002. WildBlue, in partnership with EchoStar, which owns 12% of the company, will offer consumers a bundled satellite TV service and Ka-band satellite broadband Internet service. The partners are developing a single dish and single box, which will combine the set-top box and PC modem.

It is estimated that there are 20 million to 30 million homes that will be unable to receive cable or DSL anytime soon, a natural market for satellite-delivered broadband. But high-speed Internet by satellite is more expensive for consumers-who will pay $59.99 per month for StarBand compared to a little more than $40 for cable modem or DSL service-and for the companies that launch the birds. Satellite deployment is an expensive undertaking.

Moreover, satellite Internet service is not without its own technical difficulties: It can be affected by weather conditions such as wind and lightning, by the distance signals must travel from the satellite and by too many PCs hooked up to a single dish.

The wireless-based broadband-distribution platform, MMDS, is providing Internet-access services today. Some already consider it a solid player in niche markets for broadband. Long-distance providers MCI WorldCom and Sprint together spent about $3 billion last year acquiring companies that held MMDS spectrum. Both Sprint and MCI WorldCom have DSL initiatives, which they view as a way to reach customers unlikely to receive wireline DSL.

A study conducted by investment analyst Sanford C. Bernstein and McKinsey & Co., called "Broadband!," concluded that, while MMDS has promise, it's "late to the game and will certainly be far behind cable and DSL by the time new MMDS equipment is available in 2001, at the earliest." The study also reports, "There are too many technological uncertainties to view MMDS as a likely candidate to become a near-ubiquitous third network."

The broadband study speculates that complementary broadband plays-combining a satellite channel with DSL or digital television with DSL-could find attractive niche markets in users with demand for Internet-based video or in markets where cable is slow to upgrade. More interesting, the study points out that these alliances could, over time, compete with cable's offer of video, voice and data. However, the McKinsey study says, in the long-run, cable has the edge. "Even if continued advances in VDSL [very high-speed DSL] and MPEG video technology allow telcos to offer improved video packages, we expect that hybrid fiber-coax players [cable], with their greater bandwidth and inherently more video-friendly networks, will be able to stay at least one step ahead."

Bandwidth-the next level?

But infrastructure is only part of the deployment needed to get to the next level of TV on the Internet. Bandwidth, which has quadrupled in the past four years, must continue to grow to accommodate the stringent requirements of streaming video, as well as the increasingly discriminating tastes of viewers in video quality.

These days, the fastest transmission speeds touted by cable and DSL services are in the range of 500 kilobits per second (kb/s) to 640 kb/s for downstream signals and up to 384 kb/s upstream. VHS quality is pegged at 300 kb/s, while DVD comparable video would require nearly 4 megabits, according to industry experts. With most high-speed services, connection rates of 300 kb/s to 400 kb/s are common and considered reasonable quality video over PCs. Streaming video on the Web requires roughly 175 kb/s, and television-quality over the Internet takes 750 kb/s or more. Traditional telephone modems deliver data at roughly 44 kb/s to 56 kb/s. From the FCC's point of view, high-speed services are those offering more than 200 kb/s in at least one direction.

Bit rates that deliver at least tolerable video are, indeed, here and now. But there are many factors that degrade video streams: the number of users on the network, dropped connections, noise and other kinds of traffic. Servers can also contribute to the slowdown.

Even with a cable modem or DSL, users can encounter jerky motion simply as a result of congestion on the Internet itself, not because of last-mile problems. Packets, blocks of data that carry the information necessary to be delivered to an address, are routinely dropped when there is heavy traffic on the Web and must be re-requested from the Web site before being delivered to the user's PC. With content such as e-mail or any file transfer, the arrangement of packets is not important. But with video or audio streaming, packets must be sent in the appropriate sequence or the stream will not go through.

Companies such as Cidera, PanAmSat's NET-36 venture, Akamai Technologies and iBeam Broadcasting are bypassing Internet bottlenecks using satellites to bring video streams to edge servers located closer to users.

"The Internet is just not designed to support streaming. Never was. That's the fundamental issue," says Rex Bullinger, director of broadband technology, National Cable Television Association. "The only way to fight [packet loss] is with brute force: powerful routers and networks."

However, bandwidth requirements do vary according to content. Video with lots of motion such as sports content call for higher bit rates than a newscast or a lecture. And when it comes to movies, while not always action-packed, viewers expect a certain level of quality.

Granted, 300 kb/s does not provide a television experience. But as Jupiter Research analyst David Card points out, users currently are willing to accept lower quality for content not available elsewhere, such as concerts or short films. The short form, in fact, has become popular simply because Web sites now offer the genre, for which outlets have been scarce.

The nightmare: 'Everyone gets broadband'

If the magical bit rate that delivers true TV over the Internet is upwards of 750 kb/s, then how soon will the many millions of homes be tuned in?

"That's beyond our forecast horizon," says Jupiter's Card. "Right now, real honest-to-goodness DVD quality probably requires something like a 4-Mb/s bandwidth. Nobody is deploying that."

Card ventures a guess anyway, and says the day may come in fewer than 10 years, but it will definitely be more than five years. "Let's say you do it by building out the infrastructure, putting in a lot of fiber, put a lot of content at the edge of the network. We're not even talking multicasting, which doesn't work yet either. But store movies in the local cable headend rather than in the center of the Internet. You could technically do it, but it would be very, very expensive."

Others are more optimistic. SBC's Mark Hubsher figures broadband service will be in 20 million homes within three years, an estimate that EchoStar Data Network's Jim Stratigos agrees with. Peter Negulescu of Excite @ Home thinks it could happen in less than two years.

Independent consultant and former Forrester analyst Jeremy Schwartz forecasts high-speed access will reach a "critical mass of homes by 2002 or 2003." The research company estimates penetration at 18.8 million households by 2002 and 40 million by 2005, and those numbers do not include offices and multidwelling housing. By comparison, there are 100 million TV households in the U.S. and 75 million homes with cable.

Nevertheless, Forrester's Zohar points out that massive numbers of users on the Internet could be "a huge nightmare." He believes that, when content-delivery company Akamai streamed Steve Jobs' speech to MacWorld earlier this year, nearly 6% of the Internet's capacity was consumed by users logging in for the event. Attributing the line to Akamai, Zohar says, "The worst thing that can happen to the Internet is for everyone to get broadband. Each part of the Internet needs to be bolstered, not just the last mile."

Zohar concedes that a number of companies are working on the problem with satellite solutions or terrestrial content-distribution services. But even so, he says, "the notion of getting 50 million households on concurrently is beyond our thinking right now. We forecast only about 47 million households with broadband access by 2005. I'd say it will be 2007, 2008 before we start to see that kind of world."

Despite connectivity and bandwidth impediments and the inevitable billions it will require to overcome those problems, the day will come when the PC will deliver broadcast-quality video. "It's going to happen," says NCTA's Bullinger. "Some day I expect to be sitting at home surfing channels, and I won't know-or care-whether I'm getting that video feed from a TV channel or a cable channel or a satellite or from a Web site. Some day I won't care."