ILAW: Lessig on End-to-End
Larry: My job is to take one part of what Yochai was saying and drill deeper. The significance of a principle: e2e. To introduce it I'll start with four happy stories.
1935: [] Armstrong organized a broadcast of songs from the Empire State Building to Long Island--the discovery of FM radio. Clearly a superior radio thechnology--less static than AM, much higher fidelity, could penetrate the ionosphere, broadcst at lower power, great distance.
Armstrong was working at the time for RCA. RCA controlled radio in the US. Its president was unhappy: "I didn't think he'd start up a whole damn new industry to compete with AM!"
Working to protect AM from FM. Even fought the patents, for six years. Armstrong was finally broken--and broke. RCA offered to settle, begged wife for money to do so and she refused. Armstrong stepped out of a window and killed himself.
Another "happy" story: "packet switching" technology. ATT said they hated it. It would "never work." They'd be damned if they would help him build a "competitor to ourself. "
Again, competition killed by the dominant player in the industry.
Third story: video and computers, DSL and broadband over cable. Use the Internet to stream content to consumers. In 1998 an alliance was built with Excite@Home and ATT. As a cable provider, ATT was in the business of selling content to consumers. Q: Will you allow streaming to create competitor? Answer: No. We didn't spend all this money to have our blood sucked dry.
Happy story--Kahn/Cerf--built by kids, WWW, by Cern/Swiss, ICQ--by Israelis, HotMail--by Indians, Napster, by college students. These innovations were all by kids and non-Americans.
No accident. End-to-end allows the network to open at the edges to innovation, from outsiders.
Three layers--we will address e2e layer of the logical layer. Intelligence at the ends; network as simple as possible.
Old model--ATT was network owner. New ideas could easily be crushed. "Allowable" innovation only (that which benefits themselves).
That's the old network. But in the e2e network, because there is no intelligence in the middle--you cannot discriminate. The Internet blindly, neutrally took packets and passed them along. What the user wants determines its development, detrmines innovation.
David Isenberg increasingly became a traitor to ATT--architected to be a smart network; he began to write about it within ATT (evangelizing "stupid networks"). ATT didn't like it. Eventually, he quit.
Isenberg's discovery was a rediscovery of ideas articulated by Saltzer, Clark and Reed: the end-to-end principle. The early theory in 1981 discusses functionality. If the functionality cannot completely be performed inside the network, the functionality should be placed at the edge. This was related to uses of the network, for reliability of data, security.
Not a rule, but a policy. This is a thumb on the scale, not a requirement. This is a concept familiar to lawyers. Need strong arguments to justify doing other than respecting e2e.
This has important consequences that Saltzer, Clark and Reed didn't think about. This would ramify within the competitive environment in general. [Refers to JZ's hourglass; replaces with "cooler" graphic.]
The common denominator is the IP neck of this hourglass. Tiny simple building blocks that facilitate complexity. Technical consequences and policy consequences.
First tech consequence: flexibility in how Net develops. You don't have to change things to add new flexibility at the ends. You don't have to coopt other Internet users. No coordination to build great new idea.
Voice over IP for example. You can send your voice over the Net. All you have to do is develop a program which takes voice and digitizes it, packetizes it, and spits it into protocol. You can give birth to voice over IP without talking to any network owner.
This ability to innovate without coordination = fast evolution in how the network is used. Example: Gopher (1991). It took off within 3 years. But then the WWW & $ happened. People started to charge for Gopher; no one used it anymore; they turned to the Web. Nothing anyone could do about this. Fast death.
No network owner protected Gopher. New ideas wipe out old ideas. To the extent that users like or do not like them.
e2e has features: 1/maximize competition, 2/minimize strategic threat, 3/consumer financed growth.
1/ Yochai uses the word "commons." Let's think about resources. Resources in a commons. Free in the sense tha free software and free speech is free. No one has proprietary control. Language is a resource held in common.
Negative association with "commons." Idea: Hardin's tragedy of the commons. Cattle in open range--famers keep putting more and more cows on the range. Range is destroyed. End of the resource.
This one idea completely captures how everyone thinks about resources from that moment on. But you can only have the tragedy when the resource is rivalrous. Resources which when you use it, I can't.
You take my poem by reciting it; I still have it. Language is a resources that is non-rivalrous. Contrary to there being a tragedy--there is a comedy of the commons. Hungarians have a sad story--they are the only ones who speak Hungarian. Language less valuable because it is not shared. There is no tragedy in using this resource.
As yourselves: is this a commodity that suffers a tragedy or a comedy of the commons?
People believed the Internet would die. They believed that there would be a tragedy. Reality was different than what the theory said. As more people used this resource we didn't not suffer a tragedy.
Think about something called an "innovation commons." My claim is that e2e builds an Innovation Commons. Everyone has an equal right to innovate in this space. No one is in the position to say no.
First reaction by an economist--no, there will be a tragedy. I argue to the contrary so long as e2e is preserved. The property model makes no sense in the context. It's like saying you should sell the English language.
Where you have no constraint, property systems are not doing you good, only harm.
Competitive advantage of innovation commons: one vs. many. Not one innovator. Many. Maintain principles of e2e and you increase competition.
Number two: strategic threat. In thinking about antitrust/competition law, we think about "strategic behavior." Benefit monopolist, harm consumers. Defesive monopolization--kill off new ideas.
MS Case. In 1995--Netscape/Java. A view held by some people: combo on Netscape and Java would radically transform competition.
MS hugely dominant at this time. Netscape/Java solve this "problem" by enabling people to "write once, run everywhere." Greater competition. In April 1995, Bill Gates woke up to this; he panicked. He called a meeting and said we must fight this.
The US government didn't like this. It said MS closed the platform to competition. USCA: this is illegal, unanimous.
RC, ATT and MS were protectionists.
e2e destroys the strategic opportunity. This lowers the costs of innovation.
Finally: consumer-financed growth. If you think like a utility company, you think about how you as a company will deploy a new innovation, calculate whether it will benefit you, and develop the idea over time. This is not how innovation happened in the most innovative of industries.
The Internet for example. The network evolves very quickly. 3G network vs. 802.11 wireless. 3G conceived in the old way. The 3G envisioned this amazing network. But by the time it came to be deployed, it was outclassed.
802.11 was not constrained. It is just a clue of what could happen if spectrum were further opened. Consumer-financed deployment. This is a consequence of the architecture.
End of the argument: the important thing about e2e design are the policy implications. e2e is heaven--it induces many good things. Here's the dark side: we're on our way to hell.
What we're seeing is pressure on e2e layer by physical and content layers. Corrupt the core. One way it's happening: 1) policy based routing--deciding how to treat packets, to discriminate by layering on new technology.
xBox and cable. MS has become a believer in e2e network. They were brought to this religion after a talk with cable companies over the xBox. The cable companies said, "Impress us." Pay us.
Second part: 2) the content starts to eat the conduit. Story of media concentration. Increaing concentration of message to be deployed. The point I want to make about this: important innovators in this space don't like this--and not just because of the message. Ted Turner said he would never have been able to succeed if media was as concentrated then as it is now.
FCC said: "The Internet will solve everything." But for how long? The Internet could become part of the problem.
FCC contradicts itself: "Why are you worried about maintaining e2e?"
The argument is circular.
Solutions: attempts to solve this problem. The Internet was born non-discriminatory. Plug into an outlet you get the same amount of power. Now, Internet becoming pick and choose. People are resisting in each of the three layers.
1/Open access debate; no one set of ISPs have control. Competition keep the network honest. Neutral. We in the US tried, and failed. Japan tried, and succeeded.
2/Neutral network (logical layer). FCC guarantee this will be a neutral network. MS is for this.
3/Free culture movement (content layer).
The key to facilitating e2e is guaranteeing a sliver of open and free. Lock up any of the three layers, and it jeopardizes e2e.
Any recognition of this problem in the government? No systematic understanding. There is "no need" for us to do anything. Leave it to the market to take care of it. If you leave it to the market to take care of it, it will take care of itself.
Audience Q & A:
Participant: Q about your notion of innovation commons. Does it ignore barriers of scale with innovation? Corporations doing valuable work. Not anyone can do this job.
Larry: I agree that companies innovate. Look deeper and see struggle between innovators and the business side. Business units shut down innovation [cites Innovator's Dilemma].
I disagree that more complexity is necessary. We're learning the opposite. [...]
Participant: From the point of view of the public good, I buy the e2e argument. But people will have trouble with the property aspect. There's this idea that property is sacred; how do you deal with that?
Larry: I describe this as a weird development--we have forgotten what makes property function well for society. Yochai has a paper about the evolution of free trade in India, relevant here. Would it have been better if we sold off rights to free trade? No. All sorts of places where we don't propertize, and shouldn't. We have forgotten this. The Cato Institute would like to propertize everything, to "increase efficiency." It's idiotic. It's just false. In Eldred we had economists who agreed.
Bizarre ideology with regard to property--so extreme that we're called communists if we talk about it. But free networks are about resisting extremism.
[...]
Larry: What we're talking about is balance...Yochai is now concerned about the word "commons." We're not talking either/or. We're not against property--only where it stifles innovation.