Introduction

The genesis of this essay as a letter to an editor of Wired Magazine in which I observed that there were serious errors of fact in their cover story and that the errors had consequences in reinforcing misunderstandings about what the Internet is. Unfortunately, these myths feed into a "common sense" feeling that that the Internet is an extravagance that must be contained. In fact, just the opposite is true. The reason that the Internet has contributed so much to the economy is that it is not like the old way of doing things. Unlike Television, the Internet gives all of us the opportunity to add value.

You can find the original article online in Wired Magazine. You may also want to look at an article in the New York Times that demonstrates an understanding of the difference between the Internet and the services that are available on the Web.

Alas, "Services seen as driving growth in broadband" is more typical. It posits that we have to keep the "broadband business" healthy by buying services we don't necessarily want in order to get the ability to create new services. This is completely backwards. The PC industry did not arise because of the applications created by the manufacturers. It arose because the rest of us had the opportunity to create applications.

Any plan that predicates the availability of Internet connectivity upon our acceptance of services defined by the "broadband" providers, AKA Television (and Telephony) guarantees that we cannot get improved connectivity. Such an approach is, to put it most kindly, misguided since it requires that we fund services that are at best unrelated to and at worst threatened by, the Internet. The implication is that the Internet is not valuable-enough in its own right to be funded.

The Internet is actually much more valuable than television. Requiring that the money go to more "television" isn't just misguided, it is a way to restrain the growth of the Internet and to assure that we do not get the economic benefit of the Internet.

A Letter to the Editor

I was offended by the article in the May 2001 issue of Wired magazine since it repeats some very destructive myths. Worse, it was featured on the cover. I couldn't let such a prominent collection of false statements pass without comment.

After I sent in some initial comments, the author, Charles Platt, challenged me to identify the errors in his story. I felt obliged to respond. After going through article paragraph by paragraph, the bottom line is simple. It confuses the Internet with Television and then proceeds to castigate Internet users (who are, in fact, paying their way) as freeloaders while feeling sorry for the Broadcasters and the dilemma of what do to with the all the (free) bandwidth that the FCC has given them.

The article is unfortunate in that it plays into the hands of those who demonize the Internet because their business models threatened. And the threat comes because the Internet gives us the ability to make our own choices. And that is very threatening to those who make their money by managing a scarce resource.

The resource is not scarce, they can't force us to pay their fees.

It is important to understand the Internet. One of the primary causes of the dotCom bubble was a lack of understanding on the part of investors. It wasn't that people didn't believe that businesses must earn money; it was the willingness to suspend disbelief and accept outrageous claims. The Internet seemed liked magic and everything was possible. While the Internet retains the ability to surprise us what is possible, it is important to understand how traditional businesses are challenged. The challenge might be in opportunities that provide a competitive edge. But meeting the challenge requires rethinking the fundamental assumptions of an "obvious" concept.

Television and telephony were remarkable feats of engineering. New companies that provide television and telephone services had to experiment until they found viable business models. But both of these businesses are based on charging (subscribers or advertisers) for the value, rather than the cost, of a service. The problem is that the providers prejudge what services they will provide and set the prices.

The Internet challenges this by giving each of us the ability to create our own services. You can think of the Internet as a grocery story. Instead of relying on a chef to make our meals and charge a restaurant price, we can buy the ingredients ourselves. Not only is this less expensive, we are free to create what we want.

The Internet gives us the "connectivity". We can use this to create phone calls, to view a web site, to exchange email, or watch a video stream. In fact, we can do this all at the same time. We are used to seeing each of services as prepackaged offerings provided by providers such as phone companies. While it is natural to apply this thinking to the Internet, it is very wrong. The "Internet" itself is just a transport for the services. We can define the services ourselves or choose to buy them. When we do buy a service like "Email", we are buying it from another person or company at the edge of the Internet, not inside.

Once we understand the distinction between the Internet and the services like the web, it is easy to see the problem with this article. The most exciting services will be new ones that take advantage of the Internet.

Television, as a choice of a limited number of nationwide networks doesn't make sense on the Internet since there is no reason to create such a limited set of choices. The question is not whether but when the Internet will allow causal viewing of high quality video — higher quality than current television.

Efforts to make the Internet act like television are questionable. Television is, in fact, very good television. When the Internet has the capacity to carry high quality video streams, there will be few barriers to creating new business. Trying to force the Internet to act like television prematurely gives us nice demonstrations and some snippets of video. These are interesting attempts but not, by themselves, very important to the growth of the Internet.

In fact, trying to force a television model on the Internet does serious damage. It puts those who are most threatened by the net in a position to limit its growth. And it encourages more questionable investments. It also leads to bad public policy. Strangely the article demonizes the Internet users by calling them freeloaders without any justification. Yet, when it blithely notes that the TV broadcasters are getting a free ride with current FCC policies yet sees nothing wrong. This could've been the focus of a great article. How could the editors and the author miss such an opportunity when it is presented so clearly in the stark contrast between the net users who are actually paying their way and the broadcasters who are at a loss to take advantage of the FCC's largess.

The Annotated Article

The Future Will Be Fast But Not Free
	My comments.
You want broadband. You'll get it. You'll pay for it. You'll like it.	What does the term "Broadband" mean? It used by the cable TV companies to confuse the Internet and their TV offerings. This is a poor framing for an article about the Internet.
By Charles Platt
For more than a decade, the Net successfully denied economic reality. It began, of course, as a Defense Department research network in 1969, funded entirely by the government. Even after the first commercial ISPs set up shop in the mid-'90s, the Net ignored the rules of business. Most netizens paid about the same flat rate for a valuable resource, no matter how much of it they consumed. They received the same quality of service, no matter how far from the nearest backbone they happened to live. And wherever they browsed, from Usenet to The New York Times, from the Library of Congress to Napster, they paid virtually nothing for almost anything.	What basis is there or saying that the Internet has defied economic reality? If anything, the recent rapid decrease in telecommunications prices indicates that it is the traditional telecommunications industry that has escaped reality until it was forced to catch up. And it was economic reality of the Internet that forced the issue. And it is untrue that the "netizens" paid the same flat-rate for a valuable resource. Prices vary greatly according to the type of services purchased. Just because that the number of bits isn't usually the unit of measurement nor does the charge vary by distance, doesn't mean the costs don't reflect the value of the resource. Pure connectivity is a low priced commodity like salt. Don't confuse the cost of the Internet traffic with the charges for the services offered by others via the network. If I buy a book from Amazon I pay for it. But I don't pay my ISP more because I buy a book than if I read a magazine online. This distinction between connectivity on "content" is fundamental to the economic reality of the Internet. If anything, the Internet has provided one of the best opportunities for economics systems to operate. It increases the efficiency of the marketplace and thus it is the Internet that is most in tune with economic reality. Confusing the Internet with services like Usenet and The New York Times misses the entire point of the Internet! While the Internet makes it easier to create a service like Napster, how is it any different than the old days when we'd copy a record onto a tape? Since I do pay for books on Amazon and I pay for my connectivity, it is very offensive to accuse the network users of not paying for anything.
According to traditional capitalist thinking, of course, this made no economic sense. If a product is free, businesses won't want to produce it. If a resource isn't metered, it will be abused. Yet this flat-rate environment survived and even flourished as millions of people flocked online.	This is utterly false. Does this mean that one cannot have "all you can eat" restaurants; that one must pay for a subway ride by the mile? Does this mean that every water fountain must be metered? Flat-rate pricing makes a lot of sense when the measuring is not viable. It may be that measure is expensive. At this point billing is a major, if not the largest, expense for a phone company. But the deeper problem is that there are no good metrics for charging. The providers' costs vary greatly and have no direct bearing on the user's activities. Since the flat-rate charges cover the costs of the network and charging would increases the costs, flat-rate is very rationale. What justification is there for saying that if a resource isn't metered it will be abused? It's like saying that if we allow people freedom of speech they will abuse it.
Net theory held that these peculiarities weren't peculiar at all. Technophilosophers argued that the combined effects of Moore's law and shared content had fundamentally altered the economic equation. "Bandwidth is going to be virtually free in the next era in the same way that transistors are in this era," George Gilder predicted. Stewart Brand, who first suggested that "information wants to be free," held that online content was already too cheap to meter. And John Perry Barlow suggested that traditional copyright law had become obsolete, thus exempting all digitized text, images, and music from age-old concepts of ownership. "Intellectual property law cannot be patched, retrofitted, or expanded to contain digitized expression," he wrote. "We will need to develop an entirely new set of methods as befits this entirely new set of circumstances."	The particular quotes don't necessarily represent the thinking of anyone but the author of each quote and are typically offered as starting points for discussion. It is fine to disagree with them but throwing in terms like "Technophilosophers" seems to be an ad hominem attack. Better to address the specific points. While I may not agree with all the quotes there are significant elements of truth worth exploring.
These notions were supported by technological realities. Digital files could be reproduced ad infinitum, at no additional cost, and then distributed near and far for a fraction of the cost of shipping magazines or CDs. Napster exploded into existence - a thriving realization of both Brand and Barlow's vision. As for Gilder's law that bandwidth would be free, weren't computers getting cheaper - or at least more powerful for the same price? Shouldn't the same thing happen to the hardware that enabled Net access?	"Notions"? A loaded word. But confusing Napster with the Internet represents a deep misunderstanding of the difference between the Internet and the applications. Also why are we assuming that Napster is evil? Does these mean that the record distributors are entitled to all the money the get by maintaining their exclusive channel of distribution? Seems strange that a magazine is supportive of the record companies since the magazines have to compete in a marketplace that doesn't grant them exclusivity. Why do the record companies have a lock on the distribution of music?
As it turns out, there are some good, practical reasons why not.
The most important reason is not the profits-really-do-matter epiphany that shook the dotcom world this past year. Nor is it the recent victories of the recording industry over Napster in the ongoing debate about copyright's future, or the slew of free Web services that soon after went belly up (or stopped being free). No, the immediate reason centers on a simple, straightforward technological fact: A vast fiber-optic network with countless switching nodes doesn't enjoy the same options for cost reduction as transistors. Although the data-carrying capacity of fiber has increased, it still has to be laid and maintained, and there's that old but persistent little snag of bridging the last mile to reach tens of millions of consumers. As for bandwidth, its cost shows no sign of diminishing to zero. On the contrary, higher bandwidth has meant higher monthly fees, and that trend shows every sign of continuing.	What is this arguing against? Who said profits don't matter. Why is the current record distribution system sacred? Isn't Napster just a symptom of their loss of control? Just like the Internet is challenging the ability to charge for a narrow bit stream just because it is a "phone call". And the fiber-optic network most certainly does follow Moore's law. In fact, it is catching up after being "protected" by regulations. The price/performance is improving much faster than Moore's law. The big exception is the local loop where the "content" companies (telephone and television) are able to maintain their exclusive control and they have a stake in preventing the increase in connectivity. If the local loop had the same price/performance improvement as the rest of the Internet, we would have gigabit capacity now for pennies. This is not "free" any more than a Pentium is free. But compared with the price/performance of computers of the 1980's, it is indeed free. In both cases we have a virtuous cycle in which we discover new applications and keep pushing against the limits. Instead of the improvement going to "free", we benefit from taking advantage of the new possibilities.
In short, it is time for a reality check. The landscape of the Net has changed; that cyberfrontier of the past has become a teeming city of people, transactions, and businesses. And many of the tenets of early Net thinking now seem like a shared hallucination.	Perhaps clueless investors need a reality check but the "early Net thinking" is proving to have been right on the mark. If anything, we are suffering because the thinking was too conservative and the current net addresses and mechanisms need to be scaled up.
Cheap Net access never did have much to do with new technology, anyway. It was enabled by a hidden subsidy: the old copper wires that gave digital data a free ride from your home to the nearest telephone exchange. Everyone bitched about slow connections, but this bottleneck was really a blessing in disguise. By preventing anyone from overloading the system, it allowed low, flat-rate pricing.	Cheap Net access had very much do to with new technology. And, in the US, unmeasured phone calls have been given us a big advantage over Europe where the PTTs continue to frustrate Internet usage. This has given the United States a trillion dollar advantage. Slow connections are purely an artifice of telecommunications policy. The phone network is internally digital. The lowest speed is 56KBps. Analog modems had to do very powerful processing order to achieve this speed in one direction. If ISDN were Internet-friendly, getting a 128KBps connection would not be a problem. Instead we've been forced to be more creative. First the modem allowed us to use analog phone lines and now cable modems and other strategies give us higher speed. But even these speeds are far less than what is feasible. If there are slow connections and the costs are being covered the obvious thing is to improve the infrastructure and increase capacity. The question is why has this not happened? If flat-rate pricing is so terrible, why do phone companies have such a policy for voice calls? And if I use copper that is already in the ground and requires little maintenance what is the justification for charging for usage after the installation costs have been covered? According to the author's reasoning, if I should pay for the Internet based on whether I use Usenet or The New York Times, then I should pay a different charge for "important' phone calls vs those that are not important.
All phone jacks were equal - and equally slow - because incumbent local exchange carriers wanted them that way. The ILECs saw no reason to offer upgrades that would undermine their overpriced T1s, and they didn't have to do anything they didn't want to do, because federal law protected them from competition - until the 1996 Telecommunications Act came along. It changed everything.	Yes, I agree, dial-up service is slow because the providers have been spared a marketplace that would challenge their overprices services. The 1996 yielded some changes but didn't change everything.
Suddenly, competitive local exchange carriers acquired the right to offer services like DSL, which had been developed years previously but had never been deployed by the ILECs. This was around the same time that cable TV companies finally realized the Net wasn't just for geeks anymore. Via cable modems and DSL, broadband became a reality.	DSL was developed so that the phone companies could compete for Interactive TV business - a marketplace that has been stillborn. DSL was repurposed for Internet access but has serious problems since it is a physical specification that only works in very limited situations. Remember that "Broadband" doesn't necessarily mean "Internet" access. The Cable companies' priority was finding additional sources of revenue and the idea of charging $40/month for a service that costs much less to produce was very appealing.
By the end of 2000, about 5 million US households were equipped to surf at speeds ranging from 200 Kbps (about four times as fast as a 56K modem) to 2 Mbps (40 times faster). By the end of this year, that number will almost double, according to Jupiter Research. Many frustrated DSL customers argue that the rollout should be happening faster, but it is happening.	Surfing? That is another offensive term. The net is about much more than just cruising web pages as if they were channels on TV. The Web is no longer novelty. In fact, my dishwasher backed up tonight and I was able to go to a site in Sweden and find the manual for my ten-year-old dishwasher? That's not cruising, that's doing something mundane but very productive. And it is far more efficient than trying to find my paper copy of the manual. Was I abusing the net because I didn't pay the company for looking at their manual? No more than I would have to pay if I called their customer support on the phone. In fact, I saved them a lot of money by using the Internet!
Everyone knows that the broadband era will breed a new generation of online services, but this is only half of the story. Like any innovation, broadband will inflict major changes on its environment. It will destroy, once and for all, the egalitarian vision of the Internet.	What kind of nonsense is this? Why would broadband affect an "egalitarian vision"? If anything, it should enhance it be giving people much better connectivity and let everyone be a provider of information and services.
Already, broadband has cracked the illusion that geographical location is irrelevant. In cyberspace as in realspace, location, location, and location are the most crucial factors determining quality of service. If you live in the boondocks or on the wrong side of the tracks, cable modems may not be an option at any price. If you're too far from your nearest central office, DSL may be sluggish or inoperable, because its data speeds are limited by the length of those old copper wires.	What illusion? Service has already varied according to many factors. For those without unmeasured phone service, the Internet has been expensive to use. Performance varies according to line quality and other factors. University students would benefit from networked dorms while those off campus would not. I too lament the difficulties in getting cable modem and DSL service (lame as they may be) but this is nothing new. If one looks at the economics then this is the start of an important discussion. Pulling multiple fibers into a region costs little more than pulling one so why isn't it done more often?
"If you're in a top-tier metro area, your chances are much better for having some kind of solution, whether it's a cable modem, DSL, or fixed wireless," says Beth Gage, vice president of consulting for TeleChoice, an independent telecom advisory company. So much for egalitarianism.	If anything, telecommunications using the Internet would make rural and other areas far more accessible because the economies of scale work far better with Internet protocols than traditional telecommunications since the connectivity becomes a fungible commodity. This is threatening to telecom companies since they lose their ability to charge high prices by keeping the capacity limited and then charging high prices for each service. As the Internet capacity increases, people will create the services themselves. They will pay for the full cost of the service. A phone call will be "free" only in the sense that there is no special charge for telephony bits. But those who stand to lose by such a change will try to characterize those who fend for themselves as freeloaders. In fact, it is those who charge a high price for a service like caller-ID who are the freeloaders. Telephone companies have to create special software and policies so they can block caller ID and then charge to make it available. That is freeloading!
As for pricing, here, too, broadband will change the picture radically. Flat-rate billing isn't commercially viable in an era when some users consume 1,000 times as much data as others. If you're downloading a million bits per second, the cost of those bits isn't trivial anymore; and when entertainment companies start peddling video online, pay-per-view, pay-per-byte, and pay-per-hour will be logical consequences.	This nonsense. Companies have long paid for connectivity by "pipe size" rather than the number of bits exchanged. According to this reasoning I if I pay $30/month for telephone service, I should be paying $3000/month for a video connection since it is 100 times as capable (to be conservative). Moore's law is working very well and a megabits are hardly worth measuring. Maybe at a few terabits per second things will get expensive. But not necessarily, since I could create my own terabit service between buildings and not have any incremental charge. By the article's reasoning, why does a computer than runs more than a thousand times faster than my first PC (and many times that of my first computer) cost far less?
The biggest change will be in content itself. Net snobs used to preach that the interactive medium would displace the mind-deadening experience of watching TV, but in reality, the TV industry will simply incorporate the features of the Net that appeal to it. Testifying before a House subcommittee in September, AOL Time Warner CEO Gerald M. Levin suggested that Net distribution could "enrich people's experience of television," and predicted video-on-demand.	"Net snobs"? Another ad hominem attack thrown in lieu of reasoning. The whole concept of television doesn't make any sense. If I have the capacity to casually share video streams what is the purpose of company that exists only to choose which single stream everyone in a city is supposed to watch and then charge them for the privilege? Very strange? Of course individual providers of video and music will have the option of charging for their services and content. How this will work in practice will be interesting to see as we try various experiments. But these concerns remind me of the 1960's when there was a fear that copiers would destroy the book publishing industry.
Of course, TV-like Web-video providers such as pop.com and DEN have failed spectacularly; they could hardly hope to succeed so long as they offered jerky images of atrocious quality in tiny windows.	One more reminder that the Internet is not television. And television is not the Internet. So why do people try so hard to confuse the two?
Everyone knows that the broadband era will breed a new generation of online services, but this is only half of the story. broadband will destroy, once and for all, the old egalitarian vision of the internet.
Today, online video distributed at the formerly unimaginable rate of 1 Mbps can actually look better than a VHS tape, and 1-Mbps connections are proliferating as cable providers and ILECs lay fiber closer to homes. This is the distribution system of the future: Video will become the primary broadband application when consumers realize how good it can look. This explains why AOL - always the most prescient service provider - wanted Time Warner. AOL foresaw the HBO-ification of the Net.	Why this fixation on video? Sure, it's cute, but there is more to the world than television. What indication is there of HBO-ification?
Four steps are now necessary to complete the broadband transition. First, the Net's infrastructure must be re-engineered so that it can support 1 Mbps and higher in millions of simultaneous streams. Second, new billing systems must be installed and adopted. Third, content must be compressed for economic reasons, and copy-protected to reassure its owners. Fourth, heavyweight content providers like record companies and movie studios must make their archives available online.	The net already supports higher speeds. It's the "First mile" between the home and the backbone that is the problem. What is this continuing fixation on a billing approach that has simply not worked? Perhaps the market has really chosen the current billing system because is it the one that works and usage-based billing doesn't? Or does the author not trust the marketplace? Compression? Another fixation? If the pipe sizes are increasing by a factor of 10 per year (or faster), then why accept the compromises associated with compression? Compression comes with at a price since it is lossy. MP3 is based on a very complex acoustic model and doesn't work in many cases and MP3 compressed music suffers under further processing. Compression is a useful tool but not fundamental. Copy protection is nothing to do with the Internet itself. It is a factor in determining what businesses are viable but just one. Newspapers have no protection against copying so why does one assume that copy protection is a necessity. Record companies are content distributes but the bands create the content. Owners of movie archives can decide how to manage their assets. This has little to do with the Internet itself. Just the opposite, these companies must come to terms with an Internet that is moving far faster than they are.
Upgrading the infrastructure is the biggest challenge, requiring major investment in new technology.	More important is having a marketplace structure that supports such an upgrade. Technology will follow if there is advocacy. But if those who depend on scarcity have the ability to determine the availability of connectivity, there will be little advocacy to take full advantage technologies that are already available. For example, why do the providers not add spare fibers when building new infrastructure?
Leaders in this area have not enjoyed an easy ride during the past 12 months, but ultimately they will prevail. Millions of broadband users have already voted with their wallets for high-speed access. Fast information, not free information, will drive and shape the future of the Net.	This is another offensive statement. Who is saying that "free information" is driving the net? Sure, I expect the most interesting "information" will come from sharing with others but it is only free in the sense that one benefits from participating in a society.
Step One: Fix The Pipes The Internet was designed for a few thousand people swapping plain text at 300 bps, not for millions of users sucking down trillions of bits of streaming video. DSL and cable modems have made a dent in the problem. But if everyone attaches a fat pipe to the existing backbone, it could become as overburdened as the California power grid.	The Internet was designed for much higher speeds — Ethernet was already in use — and for many people around the world. Perhaps some of the design points suffered from a failure of imagination but the basic "end to end" architecture has no such scale limits. The Internet is also very very different from the power grid. And the California problems have much more to do with policy issues than just technical and scarcity issues.
Fortunately, Internet infrastructure is upgradable and scalable. Technical improvements to enable streaming video have been engineered during the past few years by edge networks like Akamai, Speedera, Digital Island, and iBeam.	What does streaming video have to do with the Net infrastructure? Akamai et al are simply patches outside the net infrastructure and are problematic in that they posit a static asymmetric network and reward those with poor performance by allowing them to charge for local hosting.
Suppose a college student in Minnesota wants to send a stream from her dorm webcam to her family in Vermont. If she simply dumps video onto the Net through her local ISP, the stream will be broken into packets that may be routed via Chicago, Washington, DC, or New York, depending on how the Net is loaded from moment to moment. Theoretically, when the packets arrive at each destination, a media player will reassemble them in a seamless sequence; but in the real world, we know that it doesn't always work that way. A typical online video image tends to jerk and freeze, and you find yourself looking at that irritating Net Congestion message at the bottom of your RealVideo window.	This is grossly wrong!! All networks have this characteristic. With scarce capacity the latency issues become more prominent and more buffering is needed. For a unidirectional stream like television such buffering is feasible. For more interactive streams there is no substitute for having sufficient capacity. And having a virtuous cycle of innovation is the best way to get such capacity. The only reason that the current television and phone system work is that they are very much overbuilt. It's like giving each truck it's own exclusive lane between Boston and San Francisco. The fact that one can get away with such overbuilding is an indication that current services are vastly overpriced. Treating artifacts of a problematic infrastructure as fundamental is simply naive.
Now suppose the data is sent via an edge network like Akamai. The Cambridge, Massachusetts-based distribution system leases backbones (from companies like Qwest) to ensure a steady, sequential flow between the ISP where the content originates and the destination ISPs. The Akamai backbones branch and terminate in more than 8,000 edge servers in 55 countries, strategically located as close as possible to the majority of Net users. If you're in a large city or suburb, your ISP most likely receives its streaming video from an edge network, which can remove the worst of the hiccups.	Again, why this fixation on video? And why a fixation on a broadcast model that assumes we are all watching the same thing at the same time. Why not just let us each get a few megabits of connectivity and choose what we want to view or choose to use it for other purposes. Doing so is simple and viable. If one looks at a Gigabit cable TV service, why not give us each our own stream rather than broadcasting the same stream to everyone and thus creating an artificial limitation on choice.
Edge networks have been successful. The big question is whether they can stay that way.	Caching, like compression, is simply a tool that is useful, but it is far from magic and not vital when there is sufficient capacity.
"The only reason broadband works today is that not many people use it," says Steve Lerner, vice president for streaming media technology at Speedera, a company in Santa Clara, California. "Delivering 4 Mbps to every home is beyond the capability of the infrastructure in this country. You will need equipment that will not exist for another 10 years," Lerner explains, citing fiber and advanced routers.	What justification is there for this claim? The only reason few people have high capacity access is they can't get it. Why assume that just because it isn't currently available that it can't be provided? One can argue the cost of a particular solution. Those who claim it cannot be provided shouldn't be setting policies since they are clearly not advocates for connectivity.
To complicate matters, edge networks were major casualties of the past year's market dive. Akamai stock lost close to 97 percent of its value by mid-March, plunging from more than $300 a share to just over $10. Akamai's competitors fared even worse: Digital Island of San Francisco, which counts AOL, Sun, Cisco, and Microsoft among its partners and operates more than 2,300 servers in 33 countries, saw its stock plummet from more than $100 to less than $3, while its net annual loss increased from $5.3 million in 1997 to $329.9 million.	Of course "edge networks" were a major causality. They are just minor technologies that depend on the infrastructure being limited. Those who didn't understand this simply made foolish investments. If anything, this is a strong indicators that the Internet is not television and that we need better connectivity.
These numbers tell a classic ebiz story. Entrepreneurs who envisioned a streaming-video future developed new distribution technology in anticipation of the demand. But consumers proved strangely reluctant to peer at tiny flickering images, driving dozens of content providers out of business. Those that survived are scrambling for cash; their advertisers, unmoved by small audiences, are unwilling to pony up much for ineffective banner ads. And since edge networks depend on usage fees from content providers, everyone's in a short-term bind.	Again, so what? The net is not television. Those who thought it was had a learning experience.
In the future, this cash-flow problem should be resolved. High-quality, full-screen Net video will not only enable pay-per-view or pay-per-download movies, it will also introduce us to the pleasures of TV-style Net commercials that will be far more potent than banner ads. Thus, broadband will eventually generate revenue - but the system needs to bootstrap itself to get there. Most of all, distribution costs must be reduced.	OK, enough on video. It is a side issue.
According to Akamai, its typical fee for carrying data is a couple of cents per megabyte, which sounds cheap until you consider that each user receives a separate datastream. Jonathan Seelig, Akamai's cofounder and vice president of strategy and corporate development, cites a 90-minute webcast of Steve Jobs in July 2000 as one of the company's biggest projects so far. "We transmitted more than 4.3 Gbps at the peak, to 90,000 unique visitors," he says with some pride. "We delivered more than 6 terabytes. We're the only company that has proved scalability to such massive proportions."
Unfortunately, compared with the reach of magazines and TV, an audience of 90,000 isn't "massive." More important, Akamai's version of online distribution isn't cheap. If Jobs paid Akamai's usual rate, the distribution of his webcast would have put him back at least $100,000, or more than $1 per person. Worse still, any Internet content provider that tries to increase its audience must pay extra for each new viewer. Broadcast television doesn't work this way: Even if the audience for a TV show doubles, the cost of broadcasting remains the same, because one signal, from one source, travels one way, through a free medium (the air) to all users simultaneously.	Free medium? Now we are talking about freeloading!!! If it's free why can't I do my own broadcasts? Admittedly the radio spectrum should be free but it is currently managed as a scarce resource with high economic value. Broadcasts are getting the benefit of largess. And if the air is so "free", why do so many people get cable access and why are the advertising fees measured per household? Why do companies pay billions to keep others from sharing this free medium?
Edge network advocates like Digital Island's chief marketing officer, Tim Wilson, acknowledge that broadcast TV has an advantage, but they insist that as backbone capacities go up and the price of transmission goes down, Internet distribution will rule, largely because ads can be customized for narrow market segments.	Ads? Well, in some business models. But the particular choices of business models are speculative at this point. But yes, capacity will increase and video streams will be available. But most will likely be interpersonal.
Until then, alternate distribution methods will be more economical. Sunnyvale, California-based iBeam, which carries MTV videos online and boasts alliances with Dell, Microsoft, and RealNetworks, has a unique strategy of uploading its data to a satellite that transmits to edge servers, avoiding terrestrial backbones altogether. DirecPC goes a step further, beaming data directly to more than 100,000 subscribers in the US via satellite instead of piping it through the Net. However, DirecPC's rated speed of 400 Kbps is good only during off-peak hours when few users share the limited resource. Moreover, the signal is one-way; you need a dialup phone connection to carry your mouse clicks or email back to the system.	TV is TV, stop trying to turn the Internet into TV just because it's the only model some people have.
All broadband systems need upgrades before they can fulfill their promise on a mass scale, and this will require steady revenues. The good news is that the tools to monetize the datastream are here.	Utter unjustified nonsense.
While DirecPC promises that a two-way satellite link is on the way, StarBand has this up and running already. Launched in April 2000, StarBand charges $69.99 a month for Net access (and $99.99 for the premium service, with 150 TV channels courtesy of the Dish Network). StarBand-ready Compaqs are sold at RadioShack, or you can buy a special satellite modem that plugs into the USB port of your PC. Downloads are said to be as fast as 500 Kbps, while the uplink runs at 150 Kbps - although these rates may drop by more than 60 percent during busy times.	What does this have to do with the Internet? Doesn't seem to scale well.
Satellite Internet systems have a better chance of being profitable in the short term because their distribution costs aren't as closely tied to the number of users; plus, they're a boon to anyone who lives beyond the reach of other broadband sources. Their disadvantage is that they're too slow for high-quality video. Eventually they'll have to be upgraded, just like edge networks.	You invest, I certainly won't.
The bottom line: All broadband distribution systems need upgrades before they can fulfill their promise on a mass scale, and this will require steady revenues. The good news for struggling broadband startups is that the tools to monitor the transactions of this online economy are already here.	What promise? To be television?
Step Two: Streaming Revenues In December, Akamai announced MediaPlus, a software suite designed to turn bits into money. "It can help to create a pay-per-view environment," explains Seelig. "Or it can insert advertising into the datastream, customized on a geographic basis. And it enables a content provider to monetize the stream by syndicating it to other sites." Akamai is ready for the day when video looks so good online, advertisers and viewers will happily pay for it.	TV again
Geneva Technology, a 4-year-old British company that launched in the US in mid-2000, offers a similar service that runs on any wholesale-distribution system, from edge networks to large ISPs. Idar Voldnes, Geneva's president, explains that his software can process any event, such as a phone call or even a mouse click, and arrange for someone to be billed for it. "The first place we'll see this is in the mobile world," he says. "Mobile commerce will never be entirely free."	TV again
He sees the same thing happening on the desktop: "Free services will last to some extent, but businesses can't survive without revenue, and you'll have to start paying for content."	So what is new here? There are many kinds of services and business and models. This is a vacuous statement
Speedera's Steve Lerner sees the shift away from free Net services as being not just necessary, but desirable. "It's like, if everybody turns on their air conditioners, the whole power grid goes down," he says. "You have to have a penalty for using too much." Still, the payment process needn't be painful. "When you toast a bagel, it costs money, but not enough for you to think about. That's the model the Internet has to evolve to."	So Lerner doesn't understand the net either. Why keep quoting him?
A few old-guard Net idealists still insist that no one wants to pay for online content, at any price, because information wants to be free. Reacting to this, many writers, musicians, and other artists complain that they can't make money selling content online. But the artists ignore one simple fact: X-rated content isn't free but is perennially popular on the Web - amounting to a $1.5 billion industry. If people want something enough and the cost is tolerable, they will pay.	Who are these "old-guard Net idealists"? The old guard I know are very pragmatic and very savvy about economics. But once again, the author is confusing the Internet with some applications and is generalizing from a few arbitrary examples that he doesn't understand.
It's a matter of finding the price point. Music might not be as desirable as Web porn diva Danni Ashe, but it must have some value. Suppose there was a site storing a million digitized CDs, allowing up to 50 downloads for $20 a month. Wouldn't most people opt for this convenience, rather than endure the inconsistent quality and hassles of a Napster-style peer-to-peer music-sharing network?	OK, so the current music distribution system is seriously flawed and innovation is very scary. But, just like confusing the Internet with television, why assume that music is broadcast rather than shared and why assume music is just about making money?
The state of the Net today recalls the state of television in the 1970s, when cable companies started persuading consumers to pay for TV, even though it had always been free. Dish owners rebelled and indulged in massive video piracy because they were overcharged; but the piracy problem disappeared as prices dropped to a tolerable level. Today, almost everyone accepts that cable and satellite services are worth paying for, since they offer a wider range of higher-quality programming.	Once again. The Internet is not Television. Television is not the Internet.
Many businesses are betting that a similar evolution can occur online. Leading the move is a startup named iBlast, which will use existing TV stations to circumvent the costs of broadband distribution. Through a quirk in federal law, broadcasters all over America have been given free digital spectrum. Originally, they were supposed to use this for HDTV, but with amazing generosity, the FCC didn't force them to do this. Instead, broadcasters can use their extra spectrum to send a digital version of the low-quality picture format that was established as a standard back in 1953. With data compression, this picture can be squeezed into less than half a channel, leaving the rest of the new, free spectrum unused.	Free? Sounds like theft!
Broadcasters seemed unsure of how to use this surprising gift from the federal government. Then a couple of enterprising opportunists named Michael Lambert and Oliver Luckett came up with a shrewd idea. Lambert had participated in creating the Fox network, where he served as president of domestic television. Luckett had been a chief IP services architect at Qwest, the fiber-optic communications giant. Together they proposed a new nationwide TV network that would send movies, games, cartoons, and other content to broadcasters, who would then transmit these programs using their spare spectrum. Already, iBlast has partnerships with media heavyweights like Cox, Gannett, Tribune Broadcasting, the Washington Post, and the New York Times.	Yes, it is a gift. So why are the Internet users who are paying the full costs called freeloaders and the TV people who are getting a gift from the government (us) being treated with such reverence.
Because the transmissions will be digital, new receiving equipment is necessary - but you'll need it anyway as all TV broadcasts go digital over the next decade. In the beginning, the service will be pay-per-view, with 10 to 15 movies available each day. The iBlasters also talk about distributing videogames and software. You'll buy a decoder key through the company's Web site or toll-free number. After you download the content, your key will unlock it for a limited time. "We think most content providers won't want the consumer to accumulate libraries," says Luckett. "They'll make 20 movies available during a week, then pull them out and move them elsewhere, and give you 20 different movies next week."	This is getting boring. These are all highly questionable business models. Why so much focus on these particular ones?
iBlast has run test broadcasts in Los Angeles, San Diego, San Jose, Orlando, and Phoenix. Counting the 246 broadcasters that have signed up so far, iBlast will be able to reach 93 percent of US households. This gives it an immense advantage over Internet distribution - in the short term.	So what?
As terrestrial networks struggle to compete with satellite systems and local broadcasters, two factors can tip the balance in their favor: compression and copy protection. Get ready for the 500-meg feature film.
Step Three: Compress and Control As terrestrial networks struggle to compete with satellite systems and local broadcasters, two factors can tip the balance in their favor: compression and copy protection.	Nonsense. As I've said, compression is just a technical technique among many others. Nothing magic about it — just engineering and economic choices. It is a technical tradeoff. Lossy compression will become less interesting as capacity increases and people start to expect more. Remember VHS is just a compression technology that is very lossy.
...	Omitting paragraphs that continue in the vein of treating the Internet as television and related business assumptions. You can them it in the original article.
Almost all digital video is compressed before distribution. If you look closely at a picture, every few seconds you may notice the background dithering fractionally from side to side. Similarly, mouth and eye movements of talking heads may have an unnatural, incremental quality. These blemishes are introduced by MPEG-2, the world's most widely used system for squishing data. Its compression ratio ranges from the high-end 8:1, to the common 20:1, all the way to 100:1, although the flaws at that scale make the picture commercially unusable.	Yeah, so?
Now this technology standard is being challenged by Microsoft, which boasts that the compression scheme in its Media Player 8 packs down video even smaller while creating fewer artifacts. According to Dave Fester, general manager of Microsoft's digital media division, "At 750 Kbps, when we demonstrated Media Player side by side with a DVD picture, half the audience was fooled." Presumably, the other half wasn't fooled. Still, if Media Player can convince even one person that its 750-Kbps stream is comparable to a DVD, this is a major achievement. According to Fester, it means that a watchable-quality two-hour movie can fit into 500 Mbytes.	So what?
Rob Glaser, CEO of RealNetworks, claims that Microsoft is merely playing catch-up to RealPlayer 8, but it doesn't matter who's right. If a movie can be crammed into 500 megs, you can download it in about an hour via a 1-Mbps cable modem, burn it onto a CD-ROM, and share it with friends. Some people are doing this kind of thing already, using cruder video images that are illegally circulated online.
Such behavior triggers apoplectic outbursts from Hollywood's copyright-infringement junkyard dog, Jack Valenti. However, Mr. MPAA and his pack of litigators may be placated by Media Player, because it not only offers better compression, but also includes a powerful copy-protection system to assuage the anxieties of even the most paranoid movie mogul.
According to Microsoft's Fester, a movie streamed in Media Player format contains hidden codes restricting the number of times or days you will be able to view it. And you can watch it only on the computer at which you ordered the movie. If you make a copy for a friend, she'll have to pay to unlock it on her computer.
If this scheme convinces Hollywood that the Net can be a safe, lucrative environment, broadband will reach its singularity: a point where its quality ceases to be a joke and becomes a cash cow. Already, dozens of startups are positioning themselves for this bonanza.
Step Four: The Million-Movie Universe No broadband content-distribution company has done its homework more thoroughly than Intertainer, a Culver City, California, startup that has partnerships with Warner Bros., New Line, 20th Century Fox, DreamWorks SKG, Vivendi Universal, Sony Music, ESPN, PBS, and the Discovery Channel. Founded in 1996 and privately held, Intertainer raised seed money from Comcast, Intel, Microsoft, NBC, Sony, and Qwest, among others. Who could ask for more?	So, movies. We'll see what models work and what models don't. Has little to do with the Internet itself. As usual, the movie industry will be in catch up mode. But the threat presented by the Internet is minor compared with the real threat that will complete moot the studio system since digital special effects will, over time, eliminate the whole use of people as live claymation characters.
...	Omitting paragraphs that seem to ignore the many failures of "video on demand". The problem is that just saving a trip to the video store doesn't justify a massive new infrstructure. More to the point, it creates no new value. The Internet creates new value. This whole section is particular clueless in completely ignoring the lessons of the utter and complete lack of any consumer interest in these services. The users were too busy using the Internet to care.
Intertainer's high-end videos are streamed at 750 Kbps. Cincinnati Bell and Verizon offer this level of access to some consumers, while Qwest claims it's ready to provide a similar service in six cities. Intertainer CEO Jonathan Taplin says the company streams more than 500 hours of Hollywood hits, classics, TV shows, music videos, and concerts every week. You pay $3.99 to rent a relatively new release (and less for other movies), and can watch it as often as you like over a 24-hour period.
The vexing issue is whether the Hollywood players will cooperate sufficiently to attract consumers. Studios have refused to license any title to Intertainer until 40 days after it has been distributed to rental stores. "That's going to change," Taplin claims. Still, at this point you'll see a movie on an airplane before you'll find it online.
Warner reportedly is digitizing its film archives, and has negotiated with Sony for a joint video-on-demand service to be delivered over the AOL Time Warner network. Sony plans to have a new video-on-demand project dubbed MovieFly running by May - yet such online initiatives have been notoriously unfruitful. Warner Bros.' Entertaindom, launched in 1999, never received much corporate backing and was virtually abandoned a year later.
Blockbuster Video is the most recent example of a company that has failed to follow through on promises of large-scale online distribution. The company signed a distribution agreement with Enron, which has built a nationwide fiber network running servers from nCube, the Foster City, California, broadband infrastructure company. nCube's senior VP of product management, Dan Sheeran, insisted in December 2000 that the Blockbuster service would offer video-on-demand for about what you'd pay at Blockbuster's 5,000 stores. David Cox, a managing director of Enron broadband systems, was equally adamant that the Blockbuster system would work, adding, "There are other content providers we are actively courting."
But by March, Blockbuster had announced a deal with only two studios and was testing the service in just three fairly obscure markets. Courtship is futile without consummation, and Blockbuster's tentative tests in Portland, Seattle, and American Fork, Utah, fall far short of wholehearted commitment. Also, while the company has pledged to offer Universal Studios movies, it avoids mentioning how new these movies might be. At press time, there was no indication that any first-run titles, from any movie studio, will be distributed online by Blockbuster or any other service provider - and the nCube alliance had fallen apart.
So far, the most clever strategy to induce studio participation comes from SightSound Technologies, a digital movie distribution company in Mount Lebanon, Pennsylvania. SightSound claims to have rented out the first online movie, in 1999, and has always refused to use streaming video. "We have a download model, not a streaming model," says CEO Scott Sander, suggesting that many people will be happier with this, since they can use a relatively slow connection to download a movie overnight. "And, we encourage file sharing."
Digitized movies will rescue edge networks from their cash bind and speed the deployment of true high-speed Net access across the nation. And that is just the beginning of the new broadband era.
If this sounds heretical, Sander is counting on Microsoft Media Player's copy protection. When you download one of his &n for $3.95, and then make a copy for a friend, your friend must pay to unlock it - at which point SightSound gets a second $3.95 payment without having to send a second copy of the movie. In effect, he's hoping that the friends network will do his distribution for him. Also, since SightSound doesn't stream its signal, it doesn't need an impeccable connection, which eliminates the expense of an edge network. "We're basically providing an upgrade to piracy," says Sander. "Instead of trading a shaky handheld video made with a camcorder in the theater, people can trade the real thing. It looks better, and it's legal." It also costs $3.95; but for the higher quality, this may be tolerable.
Of course, if someone hacks Microsoft's encryption system, the game's over. Sander says he's not worried, because "Microsoft not only owns the source code, but the operating system, and is well able to hide the key."
Maybe so. But last year, rogue hackers circulated a compression scheme that had allegedly been "liberated" from a previous release of Microsoft Media Player. Named DivX in a satirical reference to a defunct DVD rental system, the algorithm is being used for bootleg videos online. Its originators have pledged to spread the source code among more developers than even the MPAA can sue.
Historically, copy protection has never been as secure - or as necessary - as its advocates imagine. In the 1980s, a lot of software was distributed on disks that were supposed to be copy-protected but were easily duped using bit-nibbler utilities. Eventually, software publishers were forced to abandon copy protection, although some of them predicted that their capitulation would drive them out of business. More than a decade later, their fears appear to have been groundless. Microsoft, for instance, still manages to turn a profit.
Movie studios issued equally dire warnings in 1977 when they sought an injunction to ban the Sony Betamax, fearing that videotape would bankrupt Hollywood. Today, Hollywood does quite well renting and selling recorded tapes.
The lesson should be obvious by now, but apparently it must be relearned with each new digital medium. Copy protection does help to entice content owners to offer their wares, but if the wares are overpriced, the system will be hacked. Conversely, if the product is cheap enough, protection becomes unnecessary, because most users will pay.
From this perspective, distribution over the Net needn't be a catastrophe. It could be a huge source of new revenue - provided that a wide variety of content is available.
Online, people expect more inventory, not less. If content owners try to restrict access, they'll fail to excite much interest. This is precisely what happened in disastrous test runs of video-on-demand. Compare that discredited model with the ultimate broadband scenario: Using your 45-Mbps always-on VDSL connection, you surf to your favorite movie store, where a server farm keeps a copy of almost every movie ever made, in every language and every format, including wide-screen HDTV (1,920 x 1,080 pixels). While browsing through clips, searching for an obscure 1980s Hong Kong action movie, you stumble upon a Japanese samurai film you've never heard of. Naturally, you buy them both. For $2.95 apiece, why not? In fact, the price is so low, you won't even bother to burn the movies onto DVDs. If you want to watch them again, you'll just pay for another download. As the data flows in, you have time for some video-enhanced chat via the same VDSL connection. After half an hour, when the movies have been saved onto your hard drive, your media player starts squirting bytes through your PC's $50 transmitter card, which sends its signal to the HDTV receiver in your living room. Or, you can choose to watch the movie on your 23-inch PC monitor, in sync with a friend in Hawaii. He'll maintain a text link in a separate window, swapping comments about the movie with you while it's playing.
This vision may sound as idealistic and naive as the old egalitarian model of the Net. There's a crucial difference, however: It embraces economic reality, instead of trying to deny it. In reality, cable and satellite TV services have already proved that millions of people will pay a reasonable price for a wide variety of quality programming. The same can be true with broadband, where data storage and the number of potential channels are nearly unlimited.
Movies are the most desirable, most easily digitized form of entertainment for broadband distribution. Rental or purchase fees will rescue edge networks from their cash bind and hasten the deployment of true high-speed access across the nation. But this will be just the beginning of a new era. High-speed datastreams will enable new applications that even old-style netizens could love, venturing far beyond the movie-download model.
Ten years from now, when every PalmPilot can display video, a webcam is built into every monitor, and full-screen clips are commonly sent as email attachments, the broadband metamorphosis will be complete. At that time, the egalitarian Net will be a distant memory - but no one will care. Users won't reminisce about the equalizing effect of 28.8 modems any more than car drivers yearn for a time when everyone had to drive equally slowly, because dirt roads hadn't been paved with asphalt.	Video video video? Grow up and turn off the TV once in a while! And look at your computer and see a big big world out there not filtered by the TV people.
The free ride online is over; but the ride ahead will more than compensate for anything we've lost.	Nonsense
The Net is dead. Long live the Net.	What died?
Senior writer Charles Platt (cp@panix.com) wrote about Jupiter founder Josh Harris in Wired 8.11. Copyright © 1993-2001 The Condé Nast Publications Inc. All rights reserved. Copyright © 1994-2001 Wired Digital, Inc. All rights reserved.