What It Isn’t

Look up “The Internet is” (with the quotes) on Google, and check the results. Here are the top answers, as of the day I’m writing this, in order:⁵

… terrible (theinternetisterrible.com)

… a copy machine (kk.org/thetechnium)

… a global system of interconnected computer networks that use the standard Internet Protocol Suite (TCP/IP) (Wikipedia)

…for porn (a 2005 video uploaded by Evilhoof and Flayed)

… expanding (a 2006 video uploaded by Chocolate Tampon)

… shit (theinternetisshit.com)

… made of cats (rathergood.com/cats)

… affecting traditional journalism (mashable.com)

… now sufficiently embedded in society (David Clark)

… dead (Prince)

… an agreement (Doc Searls and David Weinberger in World Of Ends.com)

… changing us (Nicholas Carr in his book The Shallows)

The answers are all over the place because the Internet is not a thing, even though it can support just about anything. So, because it can support anything, it is understood in any number of different ways. Like the universe, there are no other examples of it, and all our understandings of it are incomplete.

But, being human, we must proceed from what we do understand, starting with its protocols. These are agreements among computers and other devices about how data moves between physical networks. The Internet’s base protocols are the Transmission Control Protocol and the Internet Protocol (combined as TCP/IP). These are about as minimal as protocols can be, defining in simple terms how anything can connect to anything else, at any distance, over any available paths, and pass packets between those ends on a “best effort” basis.

By TCP/IP’s design, the Net has no single purpose—or any purpose at all—other than to move packets of data from any one end to any other end. This pushes toward zero the functional distance between those ends and minimizes the efforts required to find paths for data between them. While those efforts do have costs, the Internet itself has no interest in those costs. It has no agenda for creating or protecting scarcities, because it is not a business and therefore has no business model. It is as elemental as oxygen or a pine tree. Those have no business models, either, but are highly useful to business. So, thanks to its any-purpose design and absent business model, the Net supports trillions of dollars in business activity, much of which would not exist if the Net were not there.

Yet little in that last paragraph is obvious to most people. Instead, what’s obvious to people is that they get billed for something called “broadband” or “high-speed Internet” by their ISP, which in nearly all cases (at least in North America) is a phone or a cable TV company. Those companies don’t define the Net the same way as TCP/IP does, even though they use those protocols. Instead, ISPs define the Net as one among three services they sell as a “triple play.” The other two are telephony and cable TV.

Between those very different views, a battle line has been drawn.

Any versus Only

To simplify things a bit, look at the Net’s future as a battleground where any and only fight it out. On the side of any are the Net’s protocols. On the side of only are governments and businesses with interests in restricting and controlling access to the Net, and thwarting many purposes to which the Net might be put. This battle also happens inside our own heads, because we tend to view the Net both ways. Ironies abound.

For example, the Internet is often called a “network of networks,” yet the Net was designed to transcend the connections it employs and is therefore not reducible to them. It is not comprised of wiring and is not a “service,” even though it’s called one by ISPs.

So let’s look at the sides here. On the any side, “net-heads” (yes, they call themselves that) frame their understanding of the Net in terms of its protocols and those protocols’ virtues. On the only side, “bell-heads” (yes, they call themselves that, too) frame their understanding of the Net in terms of wiring infrastructure and billing systems.

To net-heads, the Internet is a vast new virtual space with qualities such as neutrality and generativity. To maximize economic opportunity and vitality, those virtues need to be maximized, even if phone and cable TV businesses don’t wish to acknowledge or support those virtues.

To bell-heads, the Internet’s “network of networks” is a collection of mostly private properties, with which owners should be free to do what they please. So, if what pleases them is throttling certain kinds of data traffic to maximize QoS (quality of service), too bad. They are The Market, which will grow best if they act in their own economic self-interest. Hey, look at all the good they’ve done already. (Want dial-up again, anyone?) And look at the robust competition between cable and phone companies. Isn’t that producing enough economic benefits for everybody?

Since net-heads tend to make social arguments, while bell-heads tend to make economic ones, net-heads get positioned on the left and bell-heads on the right. Between the two are boundless technical arguments that aren’t worth getting into here.

I’m a net-head, but one who wants both sides to recognize that the Net’s original design is encompassing and beneficial for economies and societies everywhere. That is, I believe the argument for the Net is the same as the one for gravity, sunlight, the periodic table, and pine trees: that it is part of nature itself. What makes the Net different from all those other products of nature is that humans made the Net for themselves.

The Net’s nature—its essential purpose—is to support everything that uses it, just as the essential purpose of a clock is to tell time. So, while the Net today relies on phone and cable connections, its support-everything purpose should not be subordinated to legacy phone and cable TV businesses. The Internet, in the neutral and generative form defined by its protocols, is a far larger and more interesting market environment than the one defined by the parochial and limited interests of phone and cable companies, both of which are desperately trying to hold on to their legacy businesses and would be better served by embracing all the opportunities the Internet opens up, for everybody.

We’re going to evolve past those old businesses anyway. Phone and cable company engineers know that, and so do many of the business leaders in those companies, even as they fight to protect their legacy businesses at all costs.

As a pro-business guy, I sympathize with phone and cable companies, which are cursed by the need to maintain margins in existing businesses while building out infrastructures that obsolete those businesses. These companies get little credit (especially from net-heads) for their genuine innovations and for their ability to innovate more. We do need them, whether we like them or not.

Likewise I sympathize with old-school ISPs: the original and surviving Internet service providers, which went into the business of doing everything the Internet Protocol encouraged, and dealing in the process with the real-world limits of wiring, spectrum, regulatory complications, and political game playing at every level. These companies never wanted to be in the TV or phone businesses, and they get little (if any) love from either side in business and policy battles.

But mostly I respect the nature of the Net as an environment that is profoundly supportive of everything that depends on it, including every kind of business that requires a data connection to the world. And I don’t think letting the carriers (or the governments) run the whole show will get us the fully market-supportive Internet we need.

Opportunity Knocking

The worldwide economy has a virtual Internet of its own, in the mesh of protocols by which sellers and buyers do business. This is one reason why business and the Internet get along so well and why the Web we know today is largely a commercial one.

Realizing that both the Internet and business are protocol-based helps us see how the Net liberates economic activities and how important it is to free the Net from the single-purpose businesses that own its “last mile.” Even phone and cable companies need to see how much business is prevented by subordinating the all-purpose Internet to their single-purpose systems—and how much more opportunity will open up for them if they help build out any-purpose Net infrastructure.

My own perspective on Net-based opportunity is anchored in the summer of 1976. I was still in my twenties, married, the father of two kids, and a former journalist and radio personality living in a small rural enclave outside Chapel Hill, North Carolina. I say former because I couldn’t get a job in journalism or radio. I had been laid off by one local radio station and turned down by the rest of them. There were no job openings at any of the area newspapers or magazines, and efforts to sell my work to publishers in New York and elsewhere had also failed. I still made a few bucks freelancing, but not enough to support my family. So I fell into working odd jobs. I even worked off rent at my landlord’s sawmill. While I eventually did well in other professions, I gave up my ambitions in journalism and radio, simply because opportunities in those fields had run out.

My view toward those opportunities was bounded by geography and employment options in North Carolina—boundaries the Internet later obliterated. I could see that obliteration coming when I first started learning about the Net in the 1980s. I experienced it firsthand when Phil Hughes, the publisher of Linux Journal, hired me to write for that magazine and a companion publication called Websmith, in 1996. That was my first job as a journalist in more than twenty years. In the decade and a half since then, I’ve continued working as an editor at Linux Journal, always following the Net closely, never ceasing to marvel at the casual geekery that made this miraculous thing possible. I’ve also been exasperated at how we mistake phone and cable company means for the Net’s actual ends.

Those ends are endless, because the Net’s capacity expands with every new protocol added to the Net’s suite of them—and because these protocols are generally compliant with the “end-to-end principle,” which was first described by Jerome H. Saltzer, David P. Reed, and David D. Clark in the 1981 paper “End-to-End Arguments in System Design.”⁶ Put simply, the end-to-end principle says the intelligence that matters most exists in boundless variety at the ends of a network, rather than with the mediating systems that sit in the middle. Therefore, they said, network protocols should be designed primarily as means for those ends, rather than to serve the parochial interests of intermediary operators, such as phone and cable companies.

In May 1997, David Isenberg, then working for AT&T, self-published a paper titled, “The Rise of the Stupid Network”—his own end-to-end argument against AT&T’s cherished belief that a network should be smart, in the sense that it needed a big all-knowing, all-doing, all-thinking company like AT&T in the middle, doing the important stuff. David wrote, “Powerful leading indicators of the Stupid Network began arriving when entrepreneurs who had no vested interest in maintaining telephone company assumptions begin to offer profitable, affordable, widely available data services.”⁷ By then, Amazon.com and eBay were two big examples. Now we have Wikipedia, Skype, Google, Twitter, Facebook, LinkedIn, and all of e-commerce.

AT&T was no doubt a smart company, and so was every other phone and cable TV company in the world. But none could begin to contemplate, much less support, any of the new and amazing things that are born and grow on the stupid Internet, every day. Why is it not obvious to everybody that the stupid Internet hosts a whole new world of opportunities?

The simple answer is that we don’t understand it. The more complex answer is that we understand it too many different ways.

Frames

We understand everything metaphorically. That is, we think and talk about everything in terms of something else. That something else is a conceptual metaphor, or what cognitive linguists also call a frame.

For example, the most common frame for time is money. That’s why we save, waste, lose, invest, and set aside time. So, think of each frame as a box of words for one subject that we borrow from when we talk about something else. Thus, we borrow the vocabulary of money when we talk about time.

Another example: life. The most common frame for life is travel. Borrowing from the travel frame, we say birth is arrival, death is departure, choices are crossroads, and careers are paths. We also say we get stuck in a rut, fall off the wagon, get lost in the woods, take the fast lane, cross that bridge when we come to it, and win the race.

The irony of metaphors is that they’re right while they’re wrong. In fact, time is not money, and life is not travel. Time is a measuring system. Life is a self-sustaining process. But, as living beings, we experience time as something valuable. We experience life as directed motion. So we talk about time in terms of money and life in terms of travel.

This goes even for our moral systems. We say good is up and bad is down because we walk upright. We say good is light and bad is dark because we are diurnal: daylight animals. If owls could talk, they might say good is dark and bad is up.⁸

I bring us down this digressive path because there are some concepts—time and life among them—that are so general and encompassing that we will never explain them perfectly, no matter how well we know them from experience. Such is the case with the Internet, which is cursed by being like nothing at all. Or, like everything.

Thus, the challenge of defining the Internet always reminds me of an old joke essay question: “Define the universe, and give three examples.”

Ending at Zero

In Small Pieces Loosely Joined, David Weinberger writes,⁹

Suppose—just suppose—that the Web is a new world that we are just beginning to inhabit. We’re like the earlier European settlers in the United States, living on the edge of the forest … Of course, while the settlers may not have known what the geography of the New World was going to be, they at least knew that there was a geography. The Web, on the other hand, has no geography, no landscape. It has no distance. It has nothing natural in it. It has few rules of behavior and fewer lines of authority. Common Sense doesn’t hold there, and uncommon sense hasn’t emerged … We don’t even know how to talk about a place that has no soil, no boundaries, no near or far.¹⁰

Craig Burton, Senior Analyst with Kuppinger Cole, combines these three ideas we’ve visited so far—end-to-end, stupid, and a new world—by describing the Internet as a vast hollow sphere, comprised entirely of ends:

I see the Net as a world we might see as a bubble. A sphere. It’s growing larger and larger, and yet inside, every point in that sphere is visible to every other one. That’s the architecture of a sphere. Nothing stands between any two points. That’s its virtue: it’s empty in the middle. The distance between any two points is functionally zero, and not just because they can see each other, but because nothing interferes with operation between any two points. There’s a word I like for what’s going on here: terraform. It’s the verb for creating a world. That’s what we’re making here: a new world.¹¹

I think of Craig’s hollow sphere as a giant three-dimensional zero. Thanks to that zero, I sat in Paris in July 2010, listening to an English language broadcast of the World Cup, from WDNC: the same North Carolina radio station whose transmitter I used to run on weekends in 1974. The cost of this wasn’t zero, just like the cost of water from a spigot isn’t zero. But the cost was low enough to ignore, and that’s a big part of what makes the Internet so appealing and useful.

Back when I worked at WDNC, we were proud that our signal reached dozens of counties in North Carolina and Virginia. It still does, but the station also reaches the whole world over the Internet, without worrying much about anybody extracting tariffs or fees between the station at one end and the listeners at many others. Thanks to the same protocol suite, I can write a blog post anywhere, put it up on a server in Cambridge, Massachusetts, and have it read by countless people all over the world, many of whom have been notified by RSS—Really Simple Syndication—without any of us belonging to any publisher’s managed syndicate.

RSS as we know it today was designed and defined by Dave Winer, who also pioneered blogging and podcasting. Today, if you look up RSS on Google, you’ll get more than 12 billion results.¹² That’s because Dave has always tried to obey one of his favorite sayings: “Ask not what the Net can do for you; ask what you can do for the Net.”¹³ (He sometimes substitutes Web for Net. Both make the same point.) The authors of other Internet protocols obey the same imperative. At this writing, Wikipedia lists forty-three protocols in its Internet protocol suite entry, and that’s not counting “more” in each of its four protocol categories.¹⁴

TCP/IP was proposed in 1974. Some protocols, such as FTP (File Transfer Protocol) are older. Others, such as HTTP (Hypertext Transfer Protocol), came along later. Most are as ownerless as a handshake or a dance move. None are encumbered by intellectual property claims, even if intellectual property is involved. Instead the Internet’s protocols embody principles that are abbreviated as “NEA.” They mean:

• Nobody owns it.
• Everybody can use it.
• Anybody can improve it.

While one could argue that the first ideal overstates or misrepresents facts, it isn’t an issue. For example, it doesn’t matter for the rest of us that somebody owns Ethernet’s patents. Those patent holders decided long ago to set Ethernet free,¹⁵ for zero-cost use and improvement by anybody (exemplifying the other two letters in NEA). One result is that almost nobody outside the networking world remembers IBM’s token ring, which competed with Ethernet back in the days when both were young and hot. Even today, some techies still regard token ring as the superior technology. But IBM wanted to sell token ring to customers, while Ethernet’s owners (Xerox, Intel, and Digital Equipment Corporation) and patent holders (notably Bob Metcalfe, widely considered the father of the protocol) wanted to make it useful to everybody, and thereby create whole new markets. So they gave it away.

The difference was between two prepositions: with and because. IBM wanted to make money with token ring, while Xerox, Intel, and Digital wanted to make money because Ethernet was free. The result was a big win for everybody, including IBM, thanks to Ethernet’s free and open nature.

Back in the late 1980s, when Craig Burton was Executive VP Corporate Marketing at Novell, I found out from a friend who worked at a component supplier that Novell was actually the world’s largest maker of Ethernet interface boards. This was when nearly everybody in business used PC clones, few of which came with Ethernet installed. That left open a market for Ethernet interface cards. Those cards then cost about a thousand dollars apiece. It was widely assumed at the time that 3Com was the king of Ethernet cards, but here I was learning that Novell was the real title holder. So I asked Craig why Novell didn’t make a big deal about its position in the Ethernet interface card market. He replied, “Because it’s going to be a zero-dollar business. In a few years, Ethernet will be a standard jack on the back of every computer.” And so it was, and still is.

Today, Craig says the same thing about Internet service. The first cost of the Net is like the first cost of Ethernet. It’s $0, or close enough. Yes, ISPs are good to have, but their model should be roads, electric service, or water distribution, not telephony or cable television. (Although they should be free to sell those services.) Your road department doesn’t say whether your driveway should be concrete or gravel. Your water department is not also your “plumbing service provider.” Yes, there are understandings about how and where you connect what’s yours to what’s theirs. But few if any of those say what you can do on your side of the relationship.

What we need here is what J. P. Rangaswami (chief scientist at Salesforce) and I call “because effects,” and economists call positive externalities. J. P. and I coined “because effects” because they are easy to explain this way: you make money because of them, rather than with them. The total “because effect” of the Internet is incalculable. The Net has become a rising tide of capacity for connection that lifts all boats—economic, social, and otherwise. That capacity rises as the inverse of the Net’s own movement toward zero in the ease and cost of connectivity.

The third NEA ideal—Anybody can improve it—is what makes the Internet such an adaptive form of infrastructure. It is also why the Net constantly improves as a marketplace, becoming more and more useful and efficient for everybody and everything that relies on it. So, while the Net can support seller-built inconveniences that limit customer choices, it can also provide customers with ways of working around those limitations.