Designing an Internet Read online

  Information Policy Series

  Edited by Sandra Braman

  The Information Policy Series publishes research on and analysis of significant problems in the field of information policy, including decisions and practices that enable or constrain information, communication, and culture irrespective of the legal siloes in which they have traditionally been located as well as state-law-society interactions. Defining information policy as all laws, regulations, and decision-making principles that affect any form of information creation, processing, flows, and use, the series includes attention to the formal decisions, decision-making processes, and entities of government; the formal and informal decisions, decision-making processes, and entities of private and public sector agents capable of constitutive effects on the nature of society; and the cultural habits and predispositions of governmentality that support and sustain government and governance. The parametric functions of information policy at the boundaries of social, informational, and technological systems are of global importance because they provide the context for all communications, interactions, and social processes.

  Virtual Economies: Design and Analysis, Vili Lehdonvirta and Edward Castronova

  Traversing Digital Babel: Information, e-Government, and Exchange, Alon Peled

  Chasing the Tape: Information Law and Policy in Capital Markets, Onnig H. Dombalagian

  Regulating the Cloud: Policy for Computing Infrastructure, edited by Christopher S. Yoo and Jean-François Blanchette

  Privacy on the Ground: Driving Corporate Behavior in the United States and Europe, Kenneth A. Bamberger and Deirdre K. Mulligan

  How Not to Network a Nation: The Uneasy History of the Soviet Internet, Benjamin Peters

  Hate Spin: The Manufacture of Religious Offense and its Threat to Democracy, Cherian George

  Big Data is Not a Monolith, edited by Cassidy R. Sugimoto, Hamid R. Ekbia, and Michael Mattioli

  Open Space: The Global Effort for Open Access to Environmental Satellite Data, Mariel Borowitz

  Decoding the Social World: Data Science and the Unintended Consequences of Communication, Sandra González-Bailón

  Designing an Internet, David D. Clark

  Designing an Internet

  David D. Clark

  The MIT Press

  Cambridge, Massachusetts

  London, England

  © 2018 Massachusetts Institute of Technology

  All rights reserved. No part of this book may be reproduced in any form or by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher.

  This book was set in Stone Serif by Westchester Publishing Services. Printed and bound in the United States of America.

  Library of Congress Cataloging-in-Publication Data

  Names: Clark, David D. (David Dana), 1944– author.

  Title: Designing an internet / David D. Clark.

  Description: Cambridge, MA: The MIT Press, 2018. | Series: Information policy series | Includes bibliographical references and indexes.

  Identifiers: LCCN 2017061377 | ISBN 9780262038607 (hardcover: alk. paper)

  Subjects: LCSH: Wide area networks (Computer networks) | Internetworking (Telecommunication) | Internet—History.

  Classification: LCC TK5105.87.C63 2018 | DDC 004.67/8—dc23

  LC record available at https://lccn.loc.gov/2017061377

  To Susan, who has supported this book with tolerance and sympathy.

  I promise not to do this again any time soon …

  Contents

  Series Editor’s Introduction

  Sandra Braman

    1  Introduction

    2  The Basics of the Internet

    3  Architecture and Design

    4  Requirements

    5  The Architecture of the Internet—A Historical Perspective

    6  Architecture and Function

    7  Alternative Network Architectures

    8  Naming and Addressing

    9  Longevity

  10  Security

  11  Availability

  12  Economics

  13  Network Management and Control

  14  Meeting the Needs of Society

  15  Looking to the Future

  Appendix: Addressing and Forwarding

  Acknowledgments

  Glossary

  Acronyms

  References

  Index

  List of Figure

  Figure 2.1   The hourglass model of the structure of the Internet, capturing the diversity of applications and technologies, connected through common agreement on the standards for IP (Internet protocol) and TCP. Adapted from Realizing the Information Future (National Research Council, 1994, 53).

  Series Editor’s Introduction

  Sandra Braman

  It is one thing to acknowledge that the Internet is a sociotechnical system, quite another to fully incorporate both the social and the technical sides into decision-making for the network and its uses. Putting politics to the side, the greatest challenge for those involved with information and communication policy today is bringing the technical and legal communities together in a common conversation. There have been technical decisions made in ignorance or disregard of legal constraints and policy problems that might or would ensue and, conversely, laws put in place by policy-makers who so little understood the technologies being regulated that the statutes or regulations were impossible to implement, bearing no relation to how the systems of concern actually work. The challenges are many, for the modes of thinking, language, and types of specific focal problems that must be addressed—and how they are framed—differ across multiple dimensions.

  David Clark’s Designing an Internet is a foundational work for those seeking to think in both social and technical terms when addressing sociotechnical policy problems. His knowledge of the Internet design process spans almost 50 years, beginning with his work on the technical issues involved in the early 1970s through his leadership of a series of recent efforts funded by the U.S. National Science Foundation that explored alternative approaches to network design issues that might be taken up in the future. The book is not only for those involved in the ever-ongoing Internet design process, but also for legal decision-makers, policy advocates and activists, entrepreneurs, and anyone who thinks about the nature of the Internet, what values are in play and how they might be in tension with each other, and where we might go from here. Designing an Internet is beautifully and accessibly written, and is so rich that many readers will find it deserving of more than one read and of keeping on the shelf as a reference work.

  The book discusses over two dozen proposals that have been put forward as alternatives to various Internet features and for how to improve the Internet going forward. Clark explains the design considerations that must be taken into account to achieve goals that include sustainability, security, access, economic viability, manageability, and meeting a wide range of social needs. Significant attention is paid to the economic, social, and political factors that provide the context within which networks are built and operate, and that will determine whether any given element of network architecture will actually succeed over time.

  Those struggling with contemporary issues, such as network neutrality, will find Clark’s explanations of Internet design enlightening; from a design perspective, to stick with this one example, the fundaments of network neutrality lie with the basic question of whether anything is known about what is in a flow or whether “bits are bits are bits.” It is with knowledge that the possibility of constraining, preventing, or violating network neu
trality is born; without it, segregating knowledge by content or service or vendor would not be possible. Technical matters such as whether a router can see the content of a packet affect not only the potential for surveillance but also economic issues of interest to content providers and third-party intermediaries. Those thinking about ways in which the nature of the state, as well as governance, is changing will also appreciate Clark’s unveiling of what has for those who are not computer scientists been hidden in technical language. While the international legal and political system is of course composed of geopolitically recognized states, for the Internet it is autonomous systems (59,000 of them at the time of writing). When faced with matters of scaling, computer scientists will typically think in hierarchical terms, while lawyers, politicians, and political scientists may well not.

  Clark’s insights into these and other differences between the technical and the social aspects of the Internet suggest the work could usefully be read hand in hand with the legal scholar Frederick Schauer’s Playing by the Rules: A Philosophical Examination of Rule-Based Decision-Making (1991) and other works examining the most basic elements of governance of structures of all kinds. His conceptualization of the expressive power of packet headers resonates with Ian Bogost’s (2007) notion of procedural rhetoric, although it operates at different levels of the sociotechnical ensemble. His discussions of specific “realizations” of the Internet bring to mind Star and Ruhleder’s (1996) discussions of the ways in which global infrastructure always and only, ultimately, exists in specific local manifestations. Social scientists would benefit by taking up the distinction between the singular steps of per-hop behaviors and the multiple steps required for actual functions when they study “effects” of the uses of technologies and of flows of information and communication. The learning can go both ways; “getting to know you” processes that generate trust within the network were learned from the social world.

  There are a number of theoretical provocations for those trying to take into account the technical in what had previously been understood as a purely social world. It goes further—Clark is thinking about aspects of communication processes or production chains that have not yet been theorized because we were not driven to do so when thinking about communication in social terms alone, such as a theory of network availability.

  Among the gifts of Designing an Internet is a succinct, and sometimes funny, history of Internet design. Defining architecture as a process, an outcome, and a discipline, Clark provides a fascinating discussion of the nature of the design process itself. Something is treated as “architecture” if agreement on an issue is necessary for the system to function, it is convenient to agree on an issue, the issue defines the basic modularity or functional dependency of the system, or it is important that the issue in question be stable over time. Core decisions about architecture affect matters of great social importance, such as the balance between surveillance and accountability, on the one hand, and anonymous action and privacy, on the other. The author helps us out with other basic concepts as well; a platform, Clark notes, is really just whatever there is that is one layer below what one is looking at.

  The Internet is working and in ever-increasing use globally, but there are lots of choices yet to be made going forward. The policy-maker’s problem is how to think about interactions between legal developments, network architecture, management of the physical network, social processes, and policy principles. The sociologist Leigh Star (1998) introduced the idea of a “Durkheim test”—as distinct from the Turing test—to evaluate whether a technical decision makes sense for humans. The computer scientist Clark agrees, arguing that we need to “allow for the future in the face of the present” (p. 43), with a “socially robust design” (p. 205).

  For computer scientists, Designing an Internet is of immediate use on the ground, contextualizing and raising the level of abstraction of issues currently being discussed within the Internet Engineering Task Force and other technical decision-making and design venues. For policy-makers, social scientists, and citizens, the book is enormously valuable as well. Clark has succeeded in providing a vocabulary, a set of concepts, and a way of thinking that can go far toward bringing the legal and technical communities into a common conversation about Internet governance.

  1   Introduction

  This is a book about how to design an internet. I say an internet rather than the Internet because the book is about not just the Internet we have today but also possible alternative conceptions of an internet—what we might instead have designed back then or might contemplate in the future. I take the word internet to describe a general-purpose, global interconnection of networks designed to facilitate communication among computers and among people using those computers. The book concerns itself with the implications of globality, the implications of generality, and the other requirements that such a network would have to meet, but it does not take the current Internet as a given—it tries to learn from the Internet of today, and from alternative proposals for what an internet might be, to draw some general conclusions and design principles about networks.

  I call these design principles architecture, so this book is about architecture as well as the specifics of an internet. There are lots of little design decisions that shape today’s Internet, but they could have been made differently and we would still have an internet. It is the basic design decisions that define the skeleton of the design, on which subsequent, more specific decisions are based. I am concerned with the question of what the essence of the design is—what defines a successful skeleton, if you will.

  This is a very personal book. It is opinionated, and I write without hesitation in the first person. It is a book-length position paper—a point of view about design. I have drawn on lots of insights from lots of people, but those people might well not agree with all of my conclusions. In this respect, the book reflects a reality of engineering, that while engineers hope that they can base their work on sound, scientific principles, engineering is also a design discipline, and design is in part a matter of taste. So what this book talks about is in part matters of taste, and if I can convince the reader about matters of taste, so much the better.

  The inspiration for this book arose out of the NSF-sponsored Future Internet Architecture program, and its predecessors, the Future Internet Design (FIND) program and the Network Science and Engineering (NetSE) program. These programs challenged the network research community to envision what an internet of 15 or 20 years from now might be, without being constrained by the Internet of today. I have been involved in this program for its duration, and I have had the chance to listen to several excellent groups of investigators discuss different approaches to designing an internet. These conversations have been very helpful in bringing into focus what is really fundamental about an internet. There have also been similar projects in other parts of the world, in particular Europe, that have contributed to my understanding. Just as one may perhaps come to understand one’s language better by studying a foreign language, one may come to understand the Internet better by studying alternative approaches. Chapter 7 provides an introduction to these various projects.

  The Internet is deeply embedded in the larger social, political, and cultural context. Assuming that we aspire to build a future global internetwork, we must accept that different parts of the world will present different contexts into which the technology must fit, so this is not a book just about technology. Indeed, technology is often not center stage. Much of the book centers on the larger issues: the economic, social, and political considerations that will determine the success or failure of a system like this, so woven into the larger world. If this book provides some insights into how the technical community can reason about this larger set of design constraints, it will have been a success from my point of view.

  My hope is that this book will be useful both to technologists concerned with network design and also a broader set of readers concerned with the character of the Internet. I
hope the book will convey a different perspective on what the Internet is, how it works, and how a range of potentially conflicting requirements have shaped its character. The book goes into detail about some critical aspects of the Internet today, including security and economics, but the deeper goal of the book is to convey how developers think about design. Networking is a subdiscipline of the field of computer science, which has its own ways of thinking about design and structuring solutions. I have tried to avoid (or else explain) some of the terms that engineers use when they talk about the Internet, but it is useful to understand some of these terms and concepts, since they are beginning to be used (often incorrectly) in nontechnical conversations about the Internet—conversations about policy, the place of the Internet in society, and the global consequences of connectivity.

  My own professional history coincides with that of the Internet. I received my PhD at MIT in 1973, the same year that the two original inventors of the Internet, Robert Kahn and Vinton Cerf, wrote their seminal paper proposing the Internet. I started working on the Internet about two years later, chaired its architecture group in the 1980s, worked on enhancing the mechanisms of the Internet in the 1990s, became more concerned with the larger sociotechnical context of the Internet in the first decade of the new millennium, and continue to do research both on technology and on the larger societal context in which the Internet sits. This book is a partial distillation of what I have learned over those 40 years.

  I begin in chapter 2 with a brief introduction to the current Internet and a review of its history from its inception in the 1970s up to today. Because the computer science community has co-opted the word architecture, I then discuss what network designers mean when they use this term. Chapter 3 talks about requirements—what it is that a network like the Internet should do. The superficial answer to that question may seem obvious—move data from point to point. However, there are many other considerations that shaped its initial design and have subsequently shaped its evolution.