Enter or Escape?:
Media Arts and Culture Online.

(Abstract from "The Architecture of Internet 2.0 by Kevin Werbach)

Kevin Werbach's article outlines attempts by large telcos to dominate and monetise the internet using anti-competitive strategies that would "wall off" sections of bandwidth and internet content. Although his example of the @Home service (and its joint venture with Excite here in Australia) has since collapsed during the buildup to the 2002 stockmarket meltdown, the ideas of capturing the market with bundled services lives on in Australia.

Telstra intends to bundle (or bring together as a "discounted" package) everything from cable TV, telephone and internet access into one tidy package that no other operator could ever hope to compete against. However, the Australian Consumer and Competition Commission (ACCC) has so far blocked their moves (at end 2002). As Werbach mentions, the internet grew because it was based on openness and access for all, and it may well be that instituting opposing ideas of closed access policies could easily kill the goose that laid the golden egg.

There are also important differences between the local and US environments. In Australia, with a smaller market, specific regulation is seen as important to protect new players and a consistent quality of Australia-wide services. In the US, open competition policy (the "let the market decide" enthusiasm of Werbach) has had some success where a number of players offer competitive services in each market.

This short abstract is quoted from "The Architecture of Internet 2.0 by Kevin Werbach" (the full version was originally published in Release 1.0, February 1999). The full article, available for a fee, after subscribing, should be at Esther Dyson's Release 1.0 site, is recommended reading, and is also very relevant to Australian communications policy. (Comments in brackets are those of the editor).

"We're approaching a critical point in the evolution of the Internet. So far we've been lucky. Back in the early 1990s, the major players in the computing and communications industries were busy chasing video on demand, closed proprietary online services and the information superhighway. They didn't much care about a bunch of academics, researchers and engineers setting up a commercial inter-network based on the Internet protocol. Many of those companies now get it... which may be cause for concern.

Architecture matters. For the most part, today's Net is open, decentralized and competitive. It fosters innovation because it is a standards-based general-purpose platform. Anyone can use it, and anyone can communicate with anyone else. Any browser (ideally) can download and display anyone's content -- and with Java or plug-ins, run their code. Companies like Mirabilis, eBay and Real Networks can develop and distribute innovative applications that spur usage, without owning any network infrastructure. Service providers must continually offer better pricing, services and support to win users' business.

The people building the next generation of high-speed access pipelines are trying to change this model. They want to tie those pipelines to the content and services they are selling, and control interconnection to the world at large. Their intractability may damage the network's openness and slow down its development. Today's Internet has grown rapidly because it is open; the next-generation Internet will not grow quickly unless it too is open and competitive. The tragedy of the cyber-commons is that competitive access benefits all providers collectively but few individually. The answer is not intrusive regulation but a set of business agreements, although government prodding may be the only way to get the parties to the table.

Anyone interested in the future of the Net should understand this debate.

Today's Net is release 1.0

Like all first commercial releases, it has bugs and missing features... but works well enough for most people. Yet in the grand scheme of things, version 1.0 is less important than what comes before and after. You can get away with more bugs and experimentation in the beta, but at some point you have to ship. On the other end, it takes several revs to fill out the reliability and feature set. Windows 1.0 wasn't terribly important except as a way station to version 3.1 and beyond....

The transition from Internet 0.9 to Internet 1.0 took place in the early 1990s, when the US National Science Foundation privatized the NSFNet backbone. Until that time the core of the Internet was a research and education network, and there were only a handful of commercial ISPs. Commercial ISPs and backbones were a necessary precursor for other developments on the Net, such as advertising-sponsored content sites, Web-driven enterprise re-engineering and electronic commerce. These innovations have driven the explosion in interest and investment in Net-related activities. However, the architecture of the network has remained largely the same in the 1990s.

The architecture of the Net will change with the move from narrowband to broadband. Right now we're witnessing version 1.5 improvements on the narrowband Net, using technologies such as load balancing, quality-of-service management and caching. Internet 2.0 will be an always-on multi-megabit network, capable of supporting voice and video streams in addition to data all the way to the end-user.

The next-generation Net does not simply offer more bandwidth; it delivers that bandwidth to the home. Bandwidth is being added constantly at the core of the network, and projects such as Internet2 are building high-speed IP networks for research and academic uses. For most users, however, the Internet experience remains the same today as five years ago. There may be more places to go and things to do, but you get there the same way. With higher bandwidth will also come always-on connectivity, enabling a whole new range of services. These developments will tie into emerging home networks.

We take it for granted that IP networks are open, but that's not preordained. Today's Net was built on the telephone network. The endpoints of that network are open in many countries thanks to government liberalization requirements. In the US, the Federal Communications Commission (FCC) stipulates that users rather than phone companies control customer premises equipment. As a result, residential users can plug modems into their phone jacks and can use any ISP they choose. This all works because the analog phone network, without modification, supports data communications up to about 50 kilobits per second (kbps).

Anything faster requires new network infrastructure. Several kinds of companies are beginning to make the necessary investments. In particular, cable and wireline local telephone companies are rolling out commercial high-speed services.

Broadband is coming! Broadband is coming!

Broadband has been just around the corner for years. Cable and telephone companies have repeatedly made grandiose promises that haven't panned out. Deployment of residential high-speed access will indeed take several years, but we can't wait to address the issue of access.

This is a critical time because the constraints that have limited the growth of high-speed services are finally disappearing. Some of the issues are technical: getting systems to the necessary price points and reliability levels for mass deployment. Although hardware was available in 1995, solutions today are much more affordable and standards-based.

The most significant drivers are competitive. The companies best able financially to make the necessary investments are incumbent cable and local telephone companies, because running new wires into homes costs so much. Even these providers must make significant investments to support high-speed Internet access, and they have no incentive to take on the cost and risk until they perceive a competitive threat.

While the highly leveraged cable companies balked at the cost of cable telephony, AT&T (read, Telstra) has the resources to make it happen. Fortunately, the same network upgrades and end-user devices (set-top boxes or standalone cable modems) will enable both high-speed Internet access and voice telephony over cable.

Cable Internet architecture

Cable Internet access platforms, differ in important ways from dial-up Internet service providers (ISPs). In effect, @Home (which has since gone belly up) is a closed network that runs on the IP protocol and interfaces with the public Internet (see Figure 1). Groups of houses are connected to shared networks; we could call them neighborhood-area networks. At a cable head-end, those networks terminate in a cable modem termination system (CMTS). The CMTS connects to the regional and national @Home backbones, which tie into the rest of the Internet at network exchange points.

Subscribers pay a monthly fee (typically $35 to $50), and as part of that fee they also get access to content from @Home and its partners. Customers can connect to any Website and can view content from other providers such as AOL, but they must pay AOL's subscription fee on top of the full @Home fee. In addition, because @Home caches content locally, its own content will have better apparent bandwidth than that of third-party content providers (caching is used by most large Australian ISPs). Because @Home makes money through advertising and commerce partnerships, the company has little incentive to provide higher-speed connectivity to outside content.

@Home controls the cable modem in the user's home and functions as the service provider. Users cannot pay a reduced fee for the high-speed pipe alone; they must purchase the @Home ISP and content offerings. Even if a user pays for another ISP's services on top of the @Home subscription fee, the primary customer relationship is still with @Home. Independent ISPs such as MindSpring and EarthLink have no control over the user's connection setup and thus cannot compete on customer service or reliability.

@Home, cable operators and AT&T claim that they need the vertically integrated access model to recoup their infrastructure investment. They say the subscription fees don't cover the full cost of the system, and they must receive advertising and transaction revenue to make the service profitable. They fiercely oppose any government order to unbundle the high-speed cable pipe from the @Home ISP and content service, arguing that government intervention would distort the free market and would chill innovation and investment.

Beyond POAS (plain old access services)

AT&T and the cable industry are fighting so hard partly because they see an opportunity to build an integrated application suite on top of the coaxial cable pipe. High-speed Internet access is merely the first and potentially the least significant offering. The local voice telephony market in the US generates roughly ten times the annual revenues of the ISP industry, and the video programming market also represents tens of billions of dollars per year. Cable operators believe they will have an easier time attacking these markets if they can build tightly coupled, application-centered networks tied to digital set-top boxes in customer homes. They are also concerned that a truly open high-speed Internet system will threaten their core video-programming revenues; @Home is required under its contracts with cable operators to limit streaming video clips over its system to 10 minutes in length.

The trouble with this vision is that it's not the Internet... and as a result it may not be as successful as the Internet. The vertically integrated model provides cable operators with better incentives to deploy facilities, but it leaves little room or incentive for third parties to develop innovative applications and services on that platform. The dynamism of the Net - the Web browser, Amazon.com, Yahoo! and so on - came about because the infrastructure was an open platform not tuned for any one kind of application. An always-on, high-speed Internet could enable many more unplanned innovations, but that will be less likely in the integrated world the cable operators are planning.

It's important to be clear here. Cable operators aren't filtering URLs to prevent customers from reaching unaffiliated content sites. The problem is that they could... and users would have no alternative. The cable operators wouldn't even have to be so blunt, because their caching architecture allows some sites to receive better treatment than others. Also, customers may not be able to use new services, such as home servers, without @Home's blessing. Any ISP faces pressures to keep customers in its own orbit, but users can normally vote with their feet.

Competitive alternatives

Open access to high-speed cable Internet systems matters only if there are not likely to be alternative open pipes into the home. If most residential customers had a choice between a vertically integrated high-speed cable Internet service and a comparable phoneline high-speed Internet service that was open to independent ISPs, there wouldn't be much of a fight about the architecture of the cable service. In fact, its very closedness would become a competitive disadvantage, like a cable service that offered a limited range of channels. Everyone agrees that the goal is competitive, unregulated markets.

Where do we go from here?

As we stated at the outset, the issue is broader than @Home, cable operators or the FCC. The question is whether an open high-speed Internet, modeled on the wildly successful Internet of today, makes economic and technical sense. We believe it does, and that efforts to supersede it with closed systems won't generate the same level of investment and innovation. In other words, a closed Internet won't grow the way the open one has. Even Microsoft was unsuccessful in building a proprietary online service with MSN.

As networking moves deeper into everyday life, last-mile infrastructure will grow in importance. Under any likely scenario, you'll still be able to send e-mail anywhere and reach any URL, but those functions will represent a shrinking fraction of the Net's functionality. If service providers continue to insist that Internet 2.0 operate on a closed model, the most significant consequence will be to delay deployment of such a network.

The market, not the government, should be able to take care of this issue. If an ISP is willing to pay a cable operator enough, the two parties should be able to reach a mutually beneficial business agreement. Instead of arguing on principles, the parties need to explore the financial trade-offs in concrete dollar values. What are the supplemental revenues cable operators could receive by tying subscribers to an affiliated ISP? What are the costs of implementing open access systems? What would the ISPs pay for using such high-speed access to reach their customers?

If the FCC tries to supervise precise details and pricing of interconnection, cable companies and their investors will justifiably think twice about building out aggressively. On the contrary, the businesspeople should hammer out agreements and tell their engineers to find technical solutions. Several ISPs should be able to sign up more customers than the cable operator alone, and both sides will benefit.

...most cable and telephone companies will continue to exclude independent ISPs from their systems. We also suspect that two to three years from now high-speed access penetration will still be lower than analysts are predicting. At that point the FCC probably will start some sort of proceeding, but it will be more difficult to open up networks that have already been deployed.

We hope we're wrong. Either way, the fork in the road is before us today. All participants in the industry should consider which option they prefer."