—7—

— at its “ends,” where users put information and applications onto the network — and that the

communications protocols themselves (the “pipes” through which information flows) be as

simple and general as possible.

17.

One consequence of this design is a principle of non-discrimination among

applications. Lower-level network layers should provide a broad range of resources that are not

particular to or optimized for any single application — even if a more efficient design for at least

some applications is thereby sacrificed. As described in a subsequent paper by Reed, Saltzer, and

Clark,

End to end arguments have … two complimentary goals: (1)
Higher-level layers, more specific to an application, are free to
(and thus expected to) organize lower level network resources to
achieve application-specific design goals efficiently (application
autonomy); (2) lower-level layers, which support many
independent applications, should provide only resources of broad
utility across applications, while providing to applications useable
means for effective sharing of resources and resolution of resource
conflicts (network transparency).3

18.

While the End-to-End design principle was first adopted for technical reasons, it

has important social and competitive features as well. End-to-end expands the competitive

horizon, by enabling a wider variety of applications to connect and use the network. It

maximizes the number of entities that can compete for the use and applications of the network.

As there is no single strategic actor who can tilt the competitive environment (the network) in

favor of itself, or no hierarchical entity that can favor some applications over others, an End-to-

End network creates a maximally competitive environment for innovation, which by design

assures competitors that they will not confront strategic network behavior.

IMAGE lem-les.doc02.gif

3See David P. Reed, Jerome H. Saltzer, and David D. Clark, Comment on Active Networking and End-to-End
Arguments
, IEEE Network 12, 3 (May/June 1998) pages 69-71.

—8—

19.

The End-to-End design of the Internet has facilitated innovation. As Reed, Saltzer

and Clark argue, for example: “had the original Internet design been optimized for telephony-

style virtual circuits (as were its contemporaries SNA and TYMNET), it would not have enabled

the experimentation that led to protocols that could support the World-Wide Web, or the flexible

interconnect that has led to the flowering of a million independent Internet Service providers.

Preserving low-cost options to innovate outside the network, while keeping the core network

services and functions simple and cheap, has been shown to have very substantial value.4

20.

The principle of End-to-End is not unique to computer networks. It has important

analogs in American constitutional law and in other legal contexts. Vis-à-vis the states, for

example, the dormant commerce clause imposes an End-to-End design on the flow of commerce:

No state is to exercise a control over the flow of commerce between states; and the kind of

control that a state may exercise over commerce flowing into that state is severely limited. The

“network” of interstate commerce is to be influenced at its ends — by the consumer and

producer — and not by intermediary actors (states) who might interfere with this flow for their

own political purposes. Vis-à-vis transportation generally, End-to-End is also how the principle

of common carriage works. The carrier is not to exercise power to discriminate in the carriage.

So long as the toll is paid, it must accept the carriage that it is offered. In both contexts, the aim

is to keep the transportation layer of intercourse simple, so as to enable the multiplication of

applications at the end.

IMAGE lem-les.doc02.gif

4 Id. at 70 (emphasis added). Note the initial ARPANET did not implement End-to-End perfectly into its design. It
was because of changes in the 1970s suggested by Vint Cerf and David P. Reed that the network we now recognize
as the Internet conformed to End-to-End.

—9—

B.

The Consequences of these Architectural Principles

21.

The effect of these Internet design principles — including, but not exclusively,

End-to-End — has been profound. By its design, the Internet has enabled an extraordinary

creativity precisely because it has pushed creativity to the ends of the network. Rather than

relying upon the creativity of a small group of innovators who work for the companies that

control the network, the End-to-End design enables anyone with an Internet connection to design

and implement a better way to use the Internet. By architecting the network to be neutral among

uses, the Internet has created a competitive environment where innovators know that their

inventions will be used if useful. By keeping the cost of innovation low, it has encouraged an

extraordinary amount of innovation.

22.

The contexts in which this innovation has occurred are many. By keeping the

network simple, and its interaction general, the Internet has facilitated the design of applications

that could not have originally been envisioned. And by keeping the cost of innovation low in the

future — especially in the context of broadband media — the Internet should continue to

facilitate innovation.

23.

End-to-end design does not only promote innovation by creating the opportunity

for innovators to offer services to the network. In our view, the effect comes as well from the

expectation that innovation will not be countered by strategic actors who might control the flow

of commerce. The potential of an actor in the distributional network to act strategically is a cost

to innovation. The expectation that an actor can act strategically is an expected cost to

innovation. Thus to the extent an actor is structurally capable of acting strategically, the rational

innovator will reckon that capacity as a cost of innovation. Compromising End-to-End will, then,

tend to undermine innovation.

[made with GoClick]