"Whole Lot O' Shakin' Goin' On!".

Internet tomorrow won't look like today

This article is the second in a two-part series. The first part appeared in the June issue of CTR.

A broadband Internet with regionalized services has serious consequences for small ISPs. Without content, hosted applications, or other value-adds, smaller providers are likely to go the way of the Dodo. Lacking the huge revenue streams (or inflated stock prices) of their larger competitors, independent ISPs will be unable to secure compelling content, and will be forced to rely on monthly fees or Web hosting, two areas which are already shrinking sources of revenue. What infrastructure analyst Jim Metzler refers to as the "AOL Time Warner Internet Century" will be a content-rich, multimedia experience that only the largest companies will be able to provide. (Indeed, a parallel shake-out is already occurring: Internet start-ups are fast losing their huge market valuations and only the sites with mega traffic and market share continue to thrive.)

In addition, a related trend is just now beginning. As the Internet has evolved, the importance of "last-mile" control has increased rapidly. After all, he who owns the wire into the home controls what goes on that wire and the ways information travels on it. When control of the curb wire is paired with control of the backbone, the business advantage is enormous. Today, only one company, AT&T, is about to have end-to-end control of its network (see below); everything else comes down to peering points and negotiated service agreements.

Keynote Systems' Lloyd Taylor believes we will see further consolidation, as backbone providers try to snatch up RBOCs or vice versa, depending on size and market share. This is the general conceit behind the AOL Time Warner Internet Century: AOL and TW content delivered through TW cable. Both companies figured out that it's better to control both sides of the connection rather than one or the other. It is also the logic behind last year's spate of mergers and acquisitions among the big backbone companies.

The same evolutionary process is likely to occur with broadband Internet access and broadband content: consolidation means more reliable content streaming, easier set-up and network maintenance, and more consistent technical support. AT&T's decision to purchase MediaOne's cable infrastructure is further proof, if any were necessary, of the value (in this case, $58 billion) of the curb wire. (At press time, a ruling by a U.S. Court of Appeal limits the number of households a single company may reach, a ruling which may affect the AT&T-MediaOne merger. The merger had not cleared the FCC or the DOJ as we went to press.)

Regulation Likely

But vertical integration presents new business and competitive issues which today seem increasingly likely to require some future form of regulation. As backbone providers sink more and more capital into their infrastructures, they will need to recoup these costs by developing fee models based on either content or usage profiles. Today's content-rich, high-traffic Web sites (think Yahoo!) demand advertising price premiums and use these premiums to develop additional content. Similarly, the most desirable content of the broadband future will be supported by both advertising and new fee-for-usage-based models, whereby backbones providers will charge one another depending on the nature and popularity of the content they exchange.

This trend will also adversely affect independent ISPs, who will increasingly be forced to shoulder more of the financial burden for the services that their customers enjoy. The willingness of consumers to absorb additional charges from their ISPs for broadband content and services is an unknown at this point.

Today, backbone providers are still in the early stages of developing models used to determine the market value of the various components that make up an Internet connection. For example, in terms of financial importance, what is the ratio of capacity to reach (scope of the network) to content to value-added services (QoS, security)? These are issues that may eventually face regulators as they try to create fair pricing models that prevent vertically integrated giants from cornering the market and inflating prices. Thus far, the FCC has shown timidity when facing mergers or acquisitions among the various cable, cellco, telco, and backbone providers; nearly everyone has been approved. In a decade the Commission may be facing enormous anti-trust problems as a result of its current policy of rubber-stamp regulation.

In addition, the Web's lack of privacy controls and abuse of consumers' rights is very likely to bring some form of government control in the future, particularly as broadband becomes widespread. In late May, the Federal Trade Commission gave up on its previous policy of allowing Internet sites to regulate themselves. Robert Pitofsky, FTC chairman, said that self-regulation, without legislation, is unlikely to provide consumers with the protection they deserve. The FTC report says, in part, that "the Commission concludes that while self-regulatory efforts have achieved some real progress, the lack of broad-based implementation of [consumer] protections online requires legislative action in order to fully protect consumers' personal information and build public confidence in electronic commerce." Several members of the Commission had dissenting opinions.

Tomorrow's Technologies

What types of technologies are likely to be used in the Internet architecture of tomorrow? Clearly, fibre is the transport medium of choice for backbones, though we can also expect to see more fibre used in LANs as the prices of components and adapters fall. DWDM has increased the carrying capacity of fibre links to such an extent that, for all practical purposes, the network will stay well ahead of the data for years to come. The bottlenecks today are in the optics of the amplifiers as well as the power output of traditional lasers. Today's edge-emitting stripe semiconductor lasers are unable to reliably and economically couple much more than 250 milliwatts into a single-mode fiber. In addition, as their name implies, edge-emitting lasers emit light from the edge of the chip rather than from its surface, creating diffusion in the beam which requires additional complexity to focus.

"One of the major manufacturing problems with edge-emitting lasers is that the chips can only be tested when they have been assembled, not when they are in wafer form," says Malcolm Thompson, CEO of Novalux, a company with new laser technology set to he released early in 2001. "In terms of manufacturing, the optical industry today is basically where the silicon industry was in the 70s and 80s," Thompson says. "The cost to performance [ratio] in the optical component industry is simply too high to allow the technologies which exist today to be effectively implemented. The underlying optical technology is here, but the laser components and manufacturing capabilities are not."

Novalux hopes to solve such problems in the Internet of tomorrow by delivering new laser technology and manufacturing processes to be used in both beam generation and amplification. The technology, called NECSEL (Novalux Extended Cavity Surface Emitting Lasers) creates a higher-power beam that is emitted in circular form from the surface of the chip, making it easier and cheaper to focus the beam without the added technical complexity of edge-emitting diode chips. NECSEL also allows chips to be tested before they are packaged, which presumably will reduce manufacturing delays and lower costs. The company is also working on new laser amplification technology which will eliminate the need for active optical switching. Current optical switching technology uses tiny mirrors to redirect laser beams from one fibre strand to another. Novalux is developing so-called "wavelength switching" which will allow light beams to be redirected without the need for moving parts.

Other, well-established companies like Lucent are partnering with smaller players like TeraBeam Networks to develop line-of-sight (also called "fibreless") optical networking components. The two companies are creating a new company, TeraBeam Internet Systems, which will create wireless, multi-gigabit networking technology for LANs and campus WANs. According to the Dell'Oro Group, in 1999 Lucent captured 34 percent (about $1.3 billion) of the $3.8 billion global DWDM equipment market. Lucent will own 30 percent of the new company, TeraBeam 70 percent.

Lucent is also working on next-generation broadband services, called the GeoVideo Network Venture, which will provide HDTV-quality, IP-based video across an international fibre network. GeoVideo Networks plans to have a presence in more than 60 markets worldwide within the next five years. Its first year rollout in this country will encompass New York, Los Angeles, Chicago, Philadelphia, and Dallas. London will be the site of its first international network hub. The service will be aimed at businesses first and later will be available to consumers. Metromedia Fiber Network will supply the metro and long-haul optical network infrastructure, and its AboveNet subsidiary will supply global IP services.

Which companies are best positioned for the new Internet infrastructure? The obvious ones are, of course, the major backbone providers: the aforementioned AT&T-MediaOne, plus Sprint/MCI-UUNet (if their merger is approved), C&W, GTE/Bell Atlantic (when the merger is complete), Digex, Qwest, and a handful of others. But other companies, particularly those that have gained a foothold in content caching services--which is basically the beginning of the move to edge services--are also well positioned. These include Akamai, CacheFlow, Digital Island/Sandpiper, Inktomi, Xcelera, and a few others, some of which may be bought by or merged with giants like Cisco and HP in the near future. The Internet Research Group predicts that the caching market will grow to an impressive $2.1 billion in 2003 from about $270 million in 1999.

In addition, a new consortium of powerful companies, founded in April, intends to push for new broadband networking architectures and content development. The Broadband Content Delivery Forum--which, though it's just a few months old, already has 100 members--had its first meeting in mid-May at Networld+Interop in Las Vegas. Nortel's Anthony Alles, the BCDF's interim chairman, says that the first meeting elected a board of directors and initiated the group's first technical conferences.

The BCDF's primary technical goal is to develop and recommend open architectures for delivering multimedia content and an improved, personalized end user experience over broadband networks. BCDF member companies include Akamai, Aliant, Alta Vista, AT&T Broadband, BT, HP, InfoSpace, Inktomi, Motorola, NBCi, Qwest, TI, and many others. More information on the forum is available at www.bcdforum.org.

Satellite Vs. Cable

Another big change in the structure of the Internet is likely to be broadband access via satellite. While much has been made of the ability of satellites to reach non-wired areas of the planet, the technology is likely to become widespread in densely populated areas first. This will be due to economies of scale, of course, but there's an additional motivator: Satellite providers are anxious to prevent DSL and cable companies from becoming entrenched in high-density markets and eroding their potential customer base.

According to recent remarks to cable industry execs by FCC Chairman William Kennard, there are currently 10 million Direct Broadcast Satellite (DBS) subscribers in the U.S., an increase of 39 percent since 1998. Recent research by Cahners In-Stat Group indicates that there are currently about 260 million households subscribing to cable TV (worldwide). By 2003, that number is expected to be 330 million. However, Cahners expects that at that time, more North American TV households will receive digital TV signals via satellite than via cable.

Indeed, with the passage of last year's Satellite Home Viewer Improvement Act, satellite providers can now offer so-called "local-into-local" service, allowing consumers to view their local network affiliates via satellite. This change has eliminated one of the cable industry's most effective selling points (and advertising campaigns). Other developments, including improved throughput and cheaper installation, are finally making satellite a viable alternative to cable. New data services are expected later this year from Hughes and Globalstar (a subsidiary of Vodaphone Airtouch); future service will be available from ICO Global and Teledesic. (The latter expects service to begin in 2004 with a network of 288 satellites.)

Dallas-based market research firm Parks Associates predicts that the combined subscriber base for broadband Internet services (including cable modem service, DSL, satellite-based Internet access, and broadband fixed wireless) will reach 24 million in the U.S. by the end of 2004, accounting for 34 percent of all online households and 50 percent of the total revenues from residential Internet access.

Storage Of The Broadband Future

Few disagree that the information age has driven data storage to the forefront of both enterprise and consumer consciousness. The disagreement comes when the discussion of the most effective storage technologies begins. There's little doubt that magnetic media will continue to play the dominant role in storage for years to come. Engineers at companies like IBM continue to increase the areal density (the amount of data that can be stored in a given amount of physical space) of magnetic platters to the point where projections indicate a few dollars per gigabyte in several years. (Disk/Trend reports that by 2002, the overall average price per megabyte for all disk drives will be 0.3 cents). The price/performance ratio of magnetic disk simply cannot be achieved by other forms of media, at least not yet.

Still, the most promising new storage technology, holography, may some day challenge magnetic media, if only for near-line and archival storage and entertainment. As a technology, holography may revolutionize the way we think about data storage: moving from linear to multilevel substrates whose capacity is for all practical purposes virtually limitless. Of course, the technology versus business issue that has plagued the optical industry since its inception isn't likely to disappear, and holography may open up an entirely new can of competing standards worms that we can only shudder to think about today.

Another area that has gotten only limited attention is miniaturization. Already, we are seeing products that once had to rely on removable flash move to miniaturized HDDs. For example, the Casio QV-3000EX Plus digital camera offer IBM's Microdrive, which provides 340MB of storage--in a camera! While flash is still easier to swap, replace, and service, miniature hard drives make a good case as the wave of the future for portable mass storage, at least until flash densities increase dramatically. The combination of higher areal density and smaller drive heads will make miniature storage devices much more common as the need for data storage expands with broadband.

Toward Tomorrow

Can the Internet continue its explosive growth without eventually being regulated? Will online taxes steal the thunder from e-commerce companies? Will old economy companies push out the startups who have dominated the Web thus far? Will viruses begin to seriously disrupt international commerce and, if so, what steps will governments and businesses take to combat them?

These and other issues will affect all of us as we adapt our business models to the Internet of tomorrow. But one thing is certain: today's Internet offers only a taste of the potential of connected computing.