Most histories of mid-twentieth century networks focus primarily on wired infrastructures, and the absence of radio in these accounts helps to promote a teleological narrative of physically connected data transfer systems progressing to lighter, cleaner networks that are increasingly disconnected from the natural world.5 Terms such as “cloud” computing point to contemporary technological discourses through which notions of dematerialized communication infrastructures travel.6 The proliferation of wireless devices and unfettered modes of data transfer means that the contemporary end user thinks more in terms of a seamless connection with digital technology rather than thinking in terms of networks and the routing of data through particular infrastructures.7 The physical means through which data travels are often hidden from the view to the contemporary end user, and contemporary scholars have called attention to what remains visible in internet infrastructure to demonstrate how institutional power is deployed through and reflected in the concrete structures of information networks.8 Despite these material conditions, networks are often represented as dematerialized and therefore separate from environmental and social practices. Fred Turner traces a history that helps explain how the postwar American engineering community reinforced these representations of networks as removed from the natural world by embracing a “New Communalist” ethic that valued information technology as a form of social progress that nevertheless involved turning away from the material conditions that undergird sociotechnical systems.9 These values can also be seen in 19th and early 20th century notions of communication technology when the telegraph, telephone, and radio were hailed as a way to triumph over previously impassible barriers to communication. Nineteenth century media helped invent the idea of communication as something fraught with the possibility of misconnections as well as a fantasy of perfect connection sometimes characterized as a union of souls.10 This utopian notion of communication neglects the physical barriers to communication, and with contemporary computer networks, distances in time and space are seemingly effaced via high speed data connections and invisible cable infrastructures.
What does the early history of wireless network design tell us about past and present notions of network materiality? The engineers who developed these designs borrowed language from nineteenth century physics to describe how the network functioned. The use of a metaphor such as “ether” suggests that these individuals sought to erase the physical components of the infrastructures they designed. In regard to the history of technology, metaphors provide a basis for the production of knowledge, not just to describe an objective reality.11 The engineer’s appropriation of the ether metaphor helped to generate the concept of networks as something removed from nature. At the same time, “ether” worked to describe networks in ways that made sense within the scientific establishment, thus reflecting Katherine Hayles’ definition of skeuomorphs. Hayles asserts that in the history of cybernetics “skeuomorphs acted as threshold devices, smoothing the transition between one conceptual constellation and another.”12 In the nascent field of computer networking, the term “ether” helped convey the way information could be sent within networks in ways that made sense to scientists at research institutions who were familiar with 19th century notions of the transfer of energy. Although “ether” was largely rejected by the scientific establishment following the publication of Einstein’s theory of special relativity, it occupies a key place in the modern history of science and also maintained cultural relevance throughout much of the 20th century.13 Analyzing the use of metaphors in network history helps to reveal the way technologies are described and understood, and the ethereal origins of LANs demonstrate a link between the scientific paradigms of the 19th century and early computer science.
Nineteenth century physicists and 20th century computer engineers share a common trait: a devotion to experimental practice and an empiricism tied to testing the physical attributes of substances in the natural world. The ether metaphor, therefore, points not just to dematerialization but also to empirical efforts to uncover a hidden material reality, and the ethereal origins of LANs demonstrate that these networks were not divorced from the physical means through which data flows. To support this claim, this paper uses interpretive discourse analysis of primary historical documents that include technical reports and engineering journal articles. The first section briefly outlines 19th and early 20th century contexts through which media were framed as dematerialized, as transcendental wishes for perfect communication. The section that follows looks to the designs and accompanying discourses of early LANs that reflect a dematerialized representation of network infrastructure. The final section establishes the connection between the immaterial representations of these networks in relation to the material circumstances of the work of early engineers. Taken together, this paper asserts that the ethereal origins of early computer networks suggest that these systems were represented as dematerialized yet at the same time were designed not to annihilate space and time but rather move data through the natural world.
Transcendental Immateriality of Nineteenth and Twentieth Century Networks
Representations of 19th and 20th century communication technologies appear dematerialized. The ostensible immateriality of contemporary communication infrastructures has roots in 19th century conceptions of the relationship between sensory experience and signal transmission.New media such as the telegraph and telephone were part of mass exhibitions at carnivals and fairs where hobbyists put on displays to showcase the spectacular abilities of these media to overcome time and space.14 Similar presentations were held for newly discovered x-rays where publics witnessed invisible transmissions that seemed to defy our expectations of visible reality.15 At these displays, spectators learned not to trust their own powers of observation thereby helping to establish a belief that one should not expect, nor is it necessary for there to be, empirically verifiable means of signal transmission. With the advent of wireless telegraphy around the turn of the 20th century, the concept of media transcending the expectations of physical reality was already well established. Radio, often termed “ether” in the early to mid-20th century, made it easier to believe in the transcendent capabilities of media to realize utopian social ideals since with radio there are no dedicated points of connection between sender and receiver. “The relationship between the wireless and ether stirred anew the old dream of ‘universal communication,’ a dream expressed in religious terms by early commentators on the telegraph.”16 The transcendental powers of media to realize the fantasy of universal communication is reflected in the metaphor of ether used to describe radio waves and other natural phenomena. Even the idea of thought transmission and telepathic signals – a particular concern amidst the burgeoning spiritualist community of the late 19th century – was often accomplished through the metaphor of “ether.”17
Ether has a long history that dates back to antiquity when it was grounded in mythology as the substance beyond the physical world where the gods lived.18 These transcendental associations were carried over to the deployment of ether in the 20th century to refer to radio waves.19 Along the way “ether” was deployed by Victorian era physicists as a term that described the transmission of electricity, heat, and light. For the ancient Greeks, ether was the substance that couldn’t be observed in the natural world, and the ether of the 19th century similarly defied empirical observation.
The transcendental hope for universal communication again resurfaced with the advent of digital information networks. One of the ways in which this fantasy was represented was through the belief that computers might overcome linguistic barriers to communication. David Gunkel likens this idealized hope for connection to the biblical fantasy of the tower of babel.20 Computers as a tool for bringing harmony across language difference reflects cybernetic notions of a distributed mind through nature via information technology. Even today this fantasy concerning the tower of babel is not uncommon. It is often deployed in regard to the advent of the TCP/IP protocols that allows connections across different local networks. Timothy Wu refers to TCP/IP as “Esperanto for machines,”21 and these protocols for connecting different networks were influenced in part by the development of the Ethernet LAN at the Xerox Palo Alto Research Center (aka Xerox PARC) in the 1970s.
Robert Metcalfe, an engineer at Xerox PARC who helped develop Ethernet, contributed to the creation of TCP/IP.22 Metcalfe advocated for having strong host protocols to facilitate universal connection across different networks.23 At the data link layer of network infrastructure, the Ethernet protocol reflects utopian aspirations for seamless communication. The protocol was designed such that two transmissions sent at the same time weren’t lost in the ether, so to speak, but rather were retransmitted after a random amount of time had elapsed thereby nearly guaranteeing successful communication within the LAN. This method for managing signal traffic within a network became commercially viable in the early 1980s following the success of the IBM PC. Ethernet became the most popular LAN system, following its standardization and intense marketing campaign. As Urs von Burg states, there was a “Babel of incompatible protocols” where there was not yet a unified LAN system.24 Ethernet became the industry standard after it was approved by the IEEE and a coalition formed across three corporations: Digital Equipment Corporation, Intel, and Xerox. The reference to babel and language barriers to problems with communication demonstrates the persistence of rhetoric in support of transcendental notions of networks in general and LANs in particular. For the Ethernet LAN system, in order to represent the system as offering a dream of universal communication there had to be protocol compatibility to link local networks (aka TCP/IP) and hardware compatibility to make it less confusing and frustrating to build local networks in the first place. Ethernet was positioned as yet another way to realize the recurring fantasy of perfect communication. As will be shown in the following section, the discourses of engineers reveal how infrastructures were positioned as immaterial and transcendent by employing 19th century conceptions of communication.
Discourses of Local Area Networks
Early wireless LAN systems demonstrate the influence of 19th century dematerialized discourse for communication technology. The ALOHAnet was an early packet-based network that inspired subsequent LAN systems. Engineer Norman Abramson looked for ways to transmit data to mainframe computers at the various University of Hawaii campuses scattered across the archipelago.25 Rather than use a costly cable-based architecture, Abramson used radio waves to transmit packets from terminal to terminal. In order for this system to function, the engineer needed to prevent data collisions, since the computers on the network shared the same frequency and one terminal might send a packet at the same time as another terminal resulting in neither packet reaching its destination. To solve this problem Abramson designed a protocol through which a terminal would resend a packet after a random interval of time had elapsed. This method decreased the likelihood of repeated packet collisions and made data transmission via radio waves possible using the shared communication medium of the electromagnetic spectrum. The ALOHAnet protocol can be likened to the Radio Act of 1912 in that both the algorithm and the policy were designed to maintain order within the shared communication medium of the electromagnetic spectrum. Since the 1912 Act mandated that all radio stations of a certain wattage in the United States be licensed, the policy helped to ensure that frequencies would be less likely to overlap with one another. In a similar fashion, the ALOHAnet protocol helped to avoid signal interference by making sure no transmission would be lost due to hindering frequencies.
The ALOHAnet protocol strongly influenced Robert Metcalfe as he designed the Xerox PARC LAN that would later be called Ethernet. Rather than use a ring topology and a token passing system architecture like the one designed at the University of California at Irvine,26 Metcalfe used a random retransmission method similar to that of the ALOHAnet.27 Metcalfe emulated the randomness of the protocol and also the fact that the system allowed for packets to be “broadcast” to all stations.28 In fact, Metcalfe traveled to Hawaii and completed his dissertation research while refining the ALOHAnet system with Abramson. In the published dissertation, titled “Packet Communication,” he espouses the virtues of broadcast-based network systems such as the ALOHAnet and the ARPANET satellite system as distinguished from point-to-point systems.29 In a memo from May 22, 1973 written to Xerox patent attorneys, Metcalfe explains the influence of the Hawaiian system. He proposes that his new design no longer be called the “Alto ALOHA Network” but instead be termed the “Ether Network.” In subsequent publications Metcalfe would often tell an origin story about how he came up with the ether metaphor as a reference to Victorian era physicists.30 And while the link to 19th century physics is essential, the ether metaphor can also be read in relation to the Ethernet protocol and the radio-based system used in Hawaii.
In addition to the discursive and technical similarities of Ethernet and ALOHAnet, both designs used wireless means for data transfer. In the latter, radio waves were the sole medium, and with Ethernet, the shared medium between stations could take the form of any number of transmission means. The “ether” to which Metcalfe refers in the 1973 memo could consist of coaxial cable, telephone lines, or radio waves. The multiple referents for the “ether” signifier remain throughout the early publication history of the technology. In fact, as long as the Ethernet architects continued to refer to “ether” as a noun (not including its use in the name “Ethernet”), designs for the LAN allowed for different types of shared media between network terminals. During the development phase, designs for the LAN even allowed for wireless media, however, the Ethernet was standardized as a system using coaxial cable. Nevertheless, the presence of radio as a transmission medium in these early schematics is reflected in the language used by engineers to describe the network.
David Boggs, co-creator of Ethernet along with Metcalfe, worked closely with radio technology. Boggs used his past experience in radio, as a hobbyist and as a side job as a technician at an NBC affiliate, to facilitate the building of the first Ethernet at the Xerox PARC.31 The conflict detection protocol of Ethernet was similar to radio-based technology to which Boggs was familiar, in particular the use of what is called “full break-in keying” among radio enthusiasts. Full break-in keying refers to the function of a transceiver which switches back to “receive” mode after sending a signal.32 The transceiver in the Ethernet system worked in a similar way. The link to radio is clear in the discourse through which Ethernet circulated during its development stage. In the first published appearance of the LAN, Metcalfe and Boggs use metaphors such as the individual “stations” that “listen” to the “ether” which remains “silent” unless it is “broadcast” to all stations.33 It would be strange to use such metaphors of audition and radio technology for a data transmission system if it weren’t clear that the engineers were influenced by the previous radio-based Hawaiian system.
The Ethernet engineers most often use “ether” to refer to the medium of communication (be it a coaxial cable, telephone line, radio waves, etc.) within a network, but at other times, “ether” points to some mysterious substance that travels within the medium. Metcalfe and Boggs write that a terminal on the network “taps into the passing Ether,” as if there is something moving within the cable.34 In a similar fashion, the engineers occasionally treat the electrical pulses that travel through the LAN as immaterial while at other times, the energy assumes a physical form analogous to natural substances. For example, Metcalfe and Boggs refer to the “pollution of the Ether,” as if it is a substance like air or water that flows through the system.35 This analogy to air is common in Victorian-era physics literature, sometimes referenced by Metcalfe. For example, when writing about the mysterious properties of ether the Victorian scientist Oliver Heaviside discussed properties of air that were previously unknown.36 The language used to describe early LANs such as Ethernet and the ALOHAnet often rely on metaphors of ephemerality. However, when these discourses are placed in relation to the material practices of the LAN architects, these systems can be seen as having an intimate connection to the natural world.
Material Practice Meets Discourses of Ephemerality
While the influence of wireless-based systems such as ALOHAnet can be seen in the ether-like language used by the engineers, at the same time the ephemeral metaphors and references that differ from the values the engineers espouse. In regard to the Ethernet architects, both Metcalfe and Boggs valued experimental practice over theory. Thus they shared values held by many Victorian-era physicists to which Metcalfe and Boggs credit as inspiration for using “ether” in the first place. For 19th century physics, ether referred to the substance for energy transmission. Of particular interest for the Ethernet engineers was the luminiferous, or “light-bearing” ether. When talking about the influence of the “ether” metaphor, the Ethernet engineers credit the Michelson-Morley experiment from 1887 in which two physicists attempted to measure the effects of the ether. This experiment, in failing to find an observable ether effect beyond an acceptable range of experimental error, called into question a wide array of theories based on the existence of a luminiferous ether.37 By referencing this experiment in particular, engineers asserted the importance of hands-on experimental practice over abstract equations.
More importantly, while building the LAN, Metcalfe and Boggs couldn’t rely solely on mathematical algorithms but instead had to test cable characteristics, device specifications, and transmission speeds.38 In fact, when Metcalfe submitted his dissertation on packet switching, his committee at Harvard rejected it on the basis that it lacked sufficient theoretical support.39 Harvard accepted the dissertation only after Metcalfe added an additional chapter on mathematically-grounded ways to improve the performance of the ALOHAnet.40 The dissertation was published not by Harvard but by the reports division of the more technically-inclined Project MAC at MIT (Metcalfe, 1973a). Metcalfe’s enthusiasm for the ether metaphor therefore positions him in relation to a long line of Victorian empiricists who valued experiment over theory. Oliver Heaviside, for example, was a contemporary of Michelson and Morley who wavered neither in his belief in the existence of the ether nor in his devotion to empiricism.41 Heaviside proclaimed to London’s Royal Society, “Mathematics is an experimental science, and definitions do not come first, but later on.”42 Heaviside’s persistence in conducting experiments concerning observable problems of energy input and maximum bandwidth led him to develop and patent the first coaxial cable, the very medium used as the “ether” for the Xerox PARC LAN.
The primary objective of engineers such as Metcalfe and Boggs was to deal with the material properties of the LAN network. Of particular concern was to keep the electric signals from degrading. This problem, referred to as signal attenuation, is crucial to engineers due to the simple fact that these systems do not annihilate space and time but rather move data through the natural world. The prototype Ethernet, the LAN developed at Xerox PARC, used coaxial cable instead of radio waves because the engineers found that cable was the best medium to combat signal attenuation.43 The cable used in Ethernet had to be passive, meaning that the medium does not produce electricity. In this regard, the transmission medium is similar to both the transmission media of the ALOHAnet and the ether models of Victorian-era physicists. The experimental practice of early LAN engineers share similar concerns regarding the medium through which energy travels. Discourses of ephemerality, exemplified by the ether metaphor, are therefore not divorced from the physical concerns of communication. The architects of the original Ethernet embraced “ether” because it was outmoded as a theoretical concept due to hands-on experimental work. The engineers chose a signifier that reflected their empirical values and active participation in creating a system that would move data through the natural world.
Now that we have uncovered the ethereal origins of local area networks we are now able to view network history in a new light. Computer engineers in the U. S. postwar period have been identified by Fred Turner as New Communalist. He writes that the New Communalists “turned away from the agonistic politics of the New Left… toward what they imagined to be a world interlinked by invisible systems.”44 New Communalists such as the engineers who developed the ALOHAnet and Ethernet LANs can be seen to demonstrate this orientation away from the visible conditions of the natural world, thereby reinforcing notions of a dematerialized network infrastructure. The emphasis on information transfer without an attendant focus on the material conditions that shape the movement of data is typical of the countercultural values of Silicon Valley engineers in the postwar period, and the ether metaphor can be seen as further evidence of the influence of the New Communalist ethos. However, the analysis of discourses for early LAN designs that employ wireless media show patterns of influence that predate the New Communalist ethos that also run contrary to the claim that these engineers were increasingly unconcerned with the material means for data transmission. The advent of computer networks builds on 19th and early 20th century conceptions of energy transfer in the case of Victorian-era ether modeling and communication infrastructure in the case of telegraphy, telephony, and radio. These ethereal origins show a history of active engagement with the physical constraints placed on energy transfer. Narratives for network history need to acknowledge that the language used to describe communication infrastructure reflects both the utopian hope for communication that transcends the natural world as well as the practical reality of sending signals across time and space. The archaeological record for early local area networks demonstrates that these systems were not progressively dematerialized but rather were produced amidst tension between notions of perfect communication and the material means for data transfer.
Janet Abbate, Inventing the Internet (Cambridge, MA: MIT Press, 1999), 114. ↩
Local area networks served as an integral part of the infrastructure that would later be termed “the Internet.” The advent of personal computers and laser printers in the 1970s facilitated the networking of small businesses. These local networks helped bring about the rapid expansion of large-scale networks as these individual networks were connected via long range cables leading to integrated wide area networks. ↩
Norman Abramson, “The ALOHA System: Another Alternative for Computer Communications,” in Proceedings of the 1970 Fall Joint Computer Conference, Volume 37 (Montvale, NJ: AFIPS Press, 1970), 281-285. ↩
Bob Metcalfe, “Ether Acquisition” (Palo Alto, CA: Xerox PARC Memorandum, 1973, May 22). ↩
Nicholas Negroponte, Being Digital (New York: Vintage, 1996); Paul E. Ceruzzi, A History of Modern Computing, 2nd edition (Cambridge, MA: MIT Press, 2003). ↩
For example, see Elizabeth Grossman, High Tech Trash: Digital Devices, Hidden Toxics, and Human Health, (Washington, DC: Island Press, 2006) chapter 1. ↩
Adrian MacKenzie, Wirelessness: Radical Empricism in Network Cultures (Cambrdige, MA: MIT Press, 2010), 8-13. ↩
For example, see Lisa Parks, “Around the Antenna Tree: The Politics of Infrastructural Visibility,” Flow (2009); Nicole Starosielski, “’Warning: Do Not Dig’: Negotiating the Visibility of Critical Infrastructures,” Journal of Visual Culture 11 (2012), 38-57. ↩
Fred Turner, From Counterculture to Cyberculture: Stewart Brand, the Whole Earth Network, and the Rise of Digital Utopianism (Chicago: University of Chicago Press, 2006), 111-118. ↩
John Durham Peters, Speaking into the Air: A History of the Idea of Communication (Chicago: University of Chicago Press, 1999). ↩
Laura C. Otis, Networking: Communication with Bodies and Machines in the Nineteenth Century (Ann Arbor, MI: University of Michigan Press, 2001), 11-13. ↩
N. Katherine Hayles, How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics (Chicago: University of Chicago Press, 1999), 17. ↩
Linda Dalrymple Henderson, “Vibratory Modernism: Boccioni, Kupka, and the Ether of Space,” in From Energy to Information Representation in Science and Technology, Art, and Literature, eds. Bruce Clarke and Linda Henderson Dalrymple (Stanford: Stanford University Press, 2002), 126-149. ↩
Paul Young, “Media on Display: A Telegraphic History of Early American Cinema,” in New Media 1740-1915, eds. Lisa Gitelman and Geoffrey B. Pingree (Cambridge, MA: MIT Press, 2003), 232-237. ↩
Simone Natale, “The Invisible Made Visible: X Rays as Attraction and Visual Medium at the End of the Nineteenth Century,” Media History 17 (2011), 345-358. ↩
Daniel Czitrom, Media and the American mind: From Morse to McLuhan (Chapel Hill: University of North Carolina Press, 1982), 65. ↩
Otis, Networking, chapter 6; Jeffrey Sconce, Haunted Media: Electronic Presence from Telegraphy to Television (Durham, NC: Duke University Press, 2000), 24-28. ↩
Joe Milutis, Ether: The Nothing That Connects Everything (Minneapolis, MN: University of Minnesota Press, 2006), chapter 1. ↩
Susan J. Douglas, Inventing American Broadcasting 1899-1922 (Baltimore: Johns Hopkins University Press, 1987),12-14. ↩
David J. Gunkel, “Lingua ex Machina: Computer-Mediated Communication and the Tower of Babel,” Configurations 7 (1999), 61-89. ↩
Timothy Wu, The Master Switch: The Rise and Fall of Information Empires (Alfred A. Knopf: New York, 2010), 196. ↩
Michael A. Hiltzik, Dealers of Lightning: Xerox PARC and the Dawn of the Computer Age (New York: Harper Collins, 1999), 293. ↩
Abbate, Inventing the Internet, 126-7. ↩
Urs von Burg, The Triumph of Ethernet: Technological Communities and the Battle for the LAN Standard (Stanford, CA: Stanford University Press, 2001), 101. ↩
Abramson, “The ALOHA System,” 282. ↩
LAN design research in the 1970s tended toward two designs: a ring topology with a token that passed between stations and a branching bus topology with a random retransmission protocol. With the ring topology structure, computers were linked in a continuous circuit through which packets were delivered. A branching bus design does not have computers linked in a ring or circle shape, and instead all computers share a connecting medium and the protocol dictates which computers may transmit via that medium at any given time. ↩
Hiltzik, Dealers of Lightning, 186-7. ↩
Robert M. Metcalfe, “How Ethernet was Invented,” IEEE Annals of the History of Computing 16 (1994), 82. ↩
Robert M. Metcalfe, Packet Communication (Cambridge, MA: Project MAC MIT, 1973), i. ↩
Robert M. Metcalfe and David R. Boggs, “Ethernet: Distributed Packet Switching for Local Computer Networks,” Communications of the ACM 19 (1976), 396; Robert M. Metcalfe, “Introduction,” in The Ethernet Sourcebook, 3rd edition (Elsevier Science Publishing Company, 1985), xi; Metcalfe, “How Ethernet was Invented,” 83. ↩
Hiltzik, Dealers of Lightning, 178. ↩
Metcalfe, “How Ethernet was Invented,” 84. ↩
Metcalfe and Boggs, “Ethernet,” 398. ↩
Metcalfe and Boggs, “Ethernet,” 396. ↩
Metcalfe and Boggs, “Ethernet,” 398. ↩
Oliver Heaviside, “On Operators in Physical Mathematics, Part II,” in Proceedings of the Royal Society of London, Vol. 54 (London: Harrison and Sons, 1893), 321. ↩
Richard Staley, Einstein’s Generation: The Origins of the Relativity Revolution (Chicago: University of Chicago Press, 2008), 60. ↩
Von Burg, Triumph of Ethernet, 73. ↩
Metcalfe, “How Ethernet was Invented,” 82; Hiltzik, Dealers of Lightning, 181. ↩
Robert M. Metcalfe, “Steady-State Analysis of a Slotted and Controlled Aloha System with Blocking,” in Proceedings of the 6th Hawaii Conference on Systems Science (North Hollywood, CA: Western Periodicals, 1973), 375-380. ↩
Paul J. Nahin, Oliver Heaviside: The Life, Work, and Times of an Electrical Genius of the Victorian Age (Baltimore, MD: The Johns Hopkins University Press, 2002), 83. ↩
Heaviside, “On Operators in Physical Mathematics,” 121. ↩
Metcalfe and Boggs, “Ethernet,” 399. ↩
Turner, From Counterculture to Cyberculture, 244. ↩
Article: Creative Commons Attribution Unported License
Image: "10,000 B.C. paragraph 2”
From: "Drawings from A Thousand Plateaus"
Original Artist: Marc Ngui
Copyright: Marc Ngui