The edge of the Internet is the place a lot of companies say they want to be. Now, investors have to sort out whether it's where they should be, too.
It's not just that several fast-growing -- yet money-losing -- companies are focusing on the edge now. What's also at stake, say some, is control over the future standards for delivery of content over the Internet -- the equivalent of
Windows operating system for getting what's on the Internet closer to the people who need it.
But first things first: How can the Internet, which has no central location, have an edge at all?
In the Internet universe, the edge is where the end-users are. For telephone dial-up subscribers to
and other Internet users, it's right next to the bank of modems to which your home computer connects via phone. For cable-modem users, it's at the head end of their local cable system.
The Efficiency Question
Efficiently getting information to the edge -- whether it's text, audio or streaming video -- becomes more important as greater numbers of people get high-speed connections to the Net. When the 28.8-bits-per-second bottleneck slowing down Net surfers is removed -- whether by cable modem or by digital subscriber line technology -- the speed at which the Internet itself delivers content becomes more apparent. And if it's too slow, Web sites fear they'll lose users to speedier ones.
"Once the last-mile problem is solved," says David Levy, senior research analyst for
, "the congestion higher up in the Internet backbone becomes more apparent."
Money's at stake, too; it costs companies money to deliver data over the Internet, so the more of it they can store at the edge, the better. It's analogous to a company in New York that wants to send out photocopies of a 20-page report to 1,000 people at an office building in San Francisco. Rather than ship 1,000 copies across the country, it's cheaper to send a single copy out to California and have it duplicated there. The economics are especially important to Internet service providers, says Rangu Salgame, CEO of self-described "edge services provider"
, because customer usage is growing faster than ISP revenue. "They're in a very tough situation," he says.
Both time and money become more relevant as the Internet slowly transforms from a text-and-image world to one carrying streams of audio and video, and from a network of PCs and other immobile computers to one that includes TVs and portable devices. "The way we build traditional content won't work," says
CEO Michael Capellas.
To get content to the edge, companies have started building content delivery networks -- a combination of computers placed around the world on-site at Internet service providers, equipped with software that efficiently delivers the Web content that end users are looking for.
The idea is that if a bunch of people are looking for popular content -- say, if a lot of customers of a California ISP want to watch the replay of a fashion show on video streamed out of New York -- the show can be sent to them on a computer sitting in San Francisco instead of one in North New Jersey.
Several companies are chasing after this great content delivery-network idea, starting with market leader
. Other firms in the publicly traded universe include
, which offers Web site-hosting along with its content-delivery network;
, a subsidiary of
holds a minority stake in Mirror Image) and
, which focuses on streaming audio, video and in-program ads.
On the privately held side are firms such as
. And then there's
, an alliance announced this summer by Internet infrastructure software company
that would create a new content delivery network with such partners as Digital Island, Exodus and privately held
, in which AOL is taking a stake.
The rewards look big. In a recent report, analyst Andrea Grosz of
First Union Securities
estimated that the content-delivery market will grow to $4.5 billion in 2005, up from $120 million in 2000. Others are more optimistic. Digital Island CEO Ruann Ernst says that content delivery will be a $15 billion to $18 billion market segment by 2003.
In the meantime, though, companies are losing money big because their fast-growing revenue can't keep up with the investments they have to make in building up the necessary infrastructure for their networks. The stocks of the publicly traded companies have plummeted this year along with the rest of the tech-stock market, since these companies are losing money as they invest heavily to build out their networks. "There's a huge amount of invested capital chasing after small revenue," says Peter Christy, a research fellow at the Jupiter Research unit of
Jupiter Media Metrix
So now several companies are trying to establish their value by striving to establish themselves as creators of a standard for distributed content and programming on the Internet.
"Think of it like an operating system," says Kieran Taylor, director of product marketing for Akamai, about his company's EdgeAdvantage technology. "In the future I think you'll see third-party services plugging into this platform as well." Digital Island hopes to be the standard for delivering streaming media.
unit, among other companies, is developing its own platform for moving content and computing out to the edge.
But the battle won't be won easily. Ed Haslam, Inktomi's chief strategist for network products, says it's harder than it looks to get people to accept a platform. He dismisses Akamai's talk of being a platform as mere marketing speak, saying the proof that something is a standard lies in whether independent software vendors are building programs on top of a company's software. And that's where Inktomi's Traffic Server product is succeeding, he says.
Christy says he thinks it will be easier for Akamai to set a standard than Inktomi, partly because of Akamai's closer contact with content providers.
But he warns that one of the challenges of staying successful as a CDN is that the technology is constantly changing. "I think you'll see continuing innovation," Christy says. "It's not fighting over a fixed or static market," he says.