Designing an Internet by David D. Clark
By David D. Clark. Cambridge, MA: MIT Press, 2018. Pp. 432. Hardcover $32.95.
Designing an Internet is an excellent, book-long case study in a key theme in the history of technology: the technologies we rely upon might well have evolved along different paths, and thinking through possible futures requires understanding the decisions (and assumptions, and coincidences) that got us to where we are. In Designing an Internet, David D. Clark walks readers through how the Internet works, alternate ways an internet might work, and the history that led to one design winning over another.
This is partly a book about history ("The design of the Internet evolved as it was reduced to practice, and its design carries its history in various decisions and the interactions among them" [p. 129]) and it takes an approach that will be familiar to historians and STS scholars ("The Internet is deeply embedded in the larger social, political, and cultural context" [p. 2] and it "does not take the current Internet as a given" [p. 1]), yet Clark is not primarily a historian of technology. Rather, he himself was a key figure in the development of the Internet, and brings a practitioner's expertise [End Page 1124] and deep familiarity with the subject matter. He cites some historical/STS literature, such as Abbate's Inventing the Internet, but Clark builds primarily upon his own experience and literature closer to his own fields.
The result, as Clark says upfront, is "a very personal book. It is opinionated, and I write without hesitation in the first person. It is a book-length position paper" (p. 1). It is "a book about how to design an Internet . . . rather than the Internet," reviewing both what happened and "what we might have designed back then or might contemplate in the future." Throughout, Clark keeps an admirable focus on the human dimensions of technology—the key role of government, for example, and the political struggles over who gets to develop standards.
The book includes occasional shades of memoir, such as in chapter five, where Clark reprints a paper he wrote in 1988 on "The Design Philosophy of the DARPA Internet Protocols," adding retrospective commentary on his earlier predictions. Historians of technology, then, can approach Designing an Internet both as primary and secondary source, and it is a valuable book in both dimensions.
The first three chapters set the groundwork: an introduction, an excellent discussion of the basics of the Internet that is approachable even to less technically-inclined readers, and an overview of key terminology (e.g. Architecture and Design).
Chapter four explores design requirement for an Internet—that is, what key questions needs answers in order to build some sort of an Internet. Most of these are later given their own full chapters. Chapter five is something of a coda, as Clark reprints and comments on his 1998 paper, as discussed above. Chapters six through fourteen each address one key issue at stake in designing an Internet: "Architecture and Function," "Alternative Network Architectures"; "Naming and Addressing"; "Longevity, Security, Availability"; "Economics, Network Management and Control"; and "Meeting the Needs of Society." In each, Clark provides an overview of the issue, usually some historical context for how the current system developed, and an examination of alternatives that have been or could be proposed. The level of technical detail in these chapters could be challenging for someone with no background in the field, though not because of the writing. Clark is admirably clear and straightforward throughout. Finally, chapter fifteen looks ahead at possible futures of the Internet.
While the chapter structure is clear, either the press or author made a questionable decision to format nearly all chapter sub-(sub-sub-) headings identically. This can lead to some confusion. For example, chapter seven's section on the "Performance" of possible alternative network architectures has twenty sub-section headings in about sixteen pages, three of which are "Discussion." Keeping track of implied sub-sections can be a minor distraction avoidable with some minor formatting differences.
In all, Designing an Internet is a fascinating, clearly argued book about [End Page 1125] the possibilities of network design, helping fuel readers' curiosity about what might have been and might yet be a very different Internet. While it was not written specifically for historians of technology, it will be an important source for any number of research programs, and no doubt its insider's perspective will make it engaging and useful for teaching both undergraduate and graduate students.
Douglas O'Reagan is the author of Taking Nazi Technology: Allied Exploitation of German Science after the Second World War. He earned a Ph.D. in history from the University of California, Berkeley, and has held academic appointments at UC Berkeley, Washington State University, and MIT.