A Prehistory of the Cloud

by Tung-Hui Hu

We may imagine the digital cloud as placeless, mute, ethereal, and unmediated. Yet the reality of the cloud is embodied in thousands of massive data centers, any one of which can use as much electricity as a midsized town. Even all these data centers are only one small part of the cloud. Behind that cloud-shaped icon on our screens is a whole universe of technologies and cultural norms, all working to keep us from noticing their existence.

Hu shows that the cloud grew out of such older networks as railroad tracks, sewer lines, and television circuits. He describes key moments in the prehistory of the cloud, from the game Spacewar as exemplar of time-sharing computers to Cold War bunkers that were later reused as data centers. Countering the popular perception of a new "cloudlike" political power that is dispersed and immaterial, Hu argues that the cloud grafts digital technologies onto older ways of exerting power over a population. But because we invest the cloud with cultural fantasies about security and participation, we fail to recognize its militarized origins and ideology. Moving between the materiality of the technology itself and its cultural rhetoric, Hu's account offers a set of new tools for rethinking the contemporary digital environment.

Genres: History, Nonfiction, Philosophy, Technology, Politics, Theory, School

Audible Audio

First published August 1, 2015

About the author

Tung-Hui Hu is a poet and a media scholar. A former network engineer, Hu is interested in understanding the hidden mechanisms within digital culture, and imagining alternatives for its future. He is an associate professor of English at the University of Michigan, and his research has been featured on BBC Radio 4, CBS News, Boston Globe, and many other venues.

His new book is Digital Lethargy: Dispatches from an Age of Disconnection, an exploration of burnout, isolation, and disempowerment in the digital underclass. Hu currently lives in Rome, where he is a 2022-23 Rome Prize Fellow in Literature at the American Academy in Rome.

Reviews

[Lee · Rating 3 out of 5]

The basic premise of this book is that the metaphors through which we understand today's internet are actually built on metaphors and infrastructure that are much older. These metaphors and infrastructure structure the way that we think about the internet, naturalizing particular ideological constructs of the internet.

I thought that this book made some interesting points, but the fact that he is conflating infrastructure and discourse (that is, metaphor) without really teasing out why these two very different things should be juxtaposed, seemed deeply troubling. Sure, the internet is built along railroad tracks and yes, the internet is built like a sewage system, but what does talking about these two things side by side help us better understand?

For much of the book, I was also wondering, "So What?" What does it matter if the internet was built along railroad tracks? What does it matter if there was this idea for a series of vans traveling around the US, downloading and uploading videos in order to get around the centralization of telelvision?

At some points, he seems to suggest the importance of his project. He says that sewage system interpolated a kind of modern subject position that was a combination of public and private, something that helped create the contemporary subject position. But, for much of the time, he does not really seem to explain why this nexus between older metaphors or older infrastructure is important.

Interesting, but it could use an editor to hammer away at it.

[David Dinaburg · Rating 5 out of 5]

For me, it started with a sticker: “There is no cloud. It’s just someone else’s computer.”

That was the first time I was confronted with the concept of “the cloud” outside of hand-waving explanations of limitless file storage. I don’t know if the structures of centralized computing are well-known or even considered by the wide majority of gmailers and instagrammers, but it is crucial to recognize that uploading something to the cloud doesn’t make it costless. A computer that stores your emails and snapchats still has to exist somewhere. A Prehistory of the Cloud gives form to that intentionally vague concept:

An icon of a cloud, the reader will recall, originally stood for any unrepresentable network on network maps, such as the Internet; in today’s computer and mobile operating systems, this cloud icon now represents a reserve of seemingly unlimited computer power, or storage space; it has become, simply, a representation of the unknown.

If one lets the cloud worry about it, one needs not be concerned by the “200 terawatt hours [of electricity] per year—roughly the same amount as South Africa—much of which is generated by fossil fuels,” that data centers currently consume. Out of sight, out of mind. And the cloud is always out of sight, after all; it is just a metaphor.

[ Read the rest of this review at DinaburgWrites.com ]

[Alva · Rating 5 out of 5]

Listened to on Audible, not read, but I definitely experienced this work as a book even so. The author is a poet, but brings not metaphorical, flowery language but rather a precision and lyric intensity to his history of the reality of cloud computing. The final section is more of a diatribe against neoliberalism than I need on my headphone-eared walks in the redwoods (being an unrepentant neoliberal myself), but I'm willing to give the author that space, in exchange for marvelous long passages about things like the early history of computing (and the original cloud) in Palo Alto and the confluence of programmers and hippie art buses. A truly amazing work -- I've looked at the author's other writing (poetry) and its scope seems much more constrained and less surprising.

[Kristine · Rating 4 out of 5]

☁️🛩💣⚒📲🗃🖤 big research interest!

[Michael Scott · Rating 2 out of 5]

TODO full review later:
+/-- Overall, a missed opportunity to capitalize on two factors: (1) the unique societal change brought by the cloud (e.g., stately and multinational corp control, but also the democratization of access to computing, data storage, and digital communication), and (2) the unique dual expertise of the author, in computer technology and in history of ideas (or, perhaps more accurate but also explaining the result, new-media critique).
+++ Excellent short history of the main technology of the cloud, from infrastructure, through networking, to the software layer. This history also includes key characters such as Turing-award-winner Corbató and father-of-the-Intergalactic-Internet Licklider (but misses father-of-queueing-theory Leonard Kleinrock and father-of-the-grid Ian Foster and... well, quite a few others), and very interesting links to pop culture, US military, and US government procedures that anticipate the fears about data privacy and reliance on networks/computers of today. Very well and concisely done.
--- Unfortunately, there is also the non-technical side, where the author launches into a dark, highly speculative, wordy exposition of anecdotal situations where the cloud can turn into a nefarious force, or is perverting our society. Maybe if this part would have had more balance, it would have been more interesting as an objective account; as it stands, it looks like the author is just trying to advance his career through biased argumentation that just happens to match his academic community's views. A few alternatives: the cloud currently offers access to services previously affordable only for a few; the cloud allows research and engineering labs to explore new solutions to prohlems they could previously not afford to even consider; the cloud allows creatives to develop and ahowcase their portfolios and products, for everyone to see; the cloud allows hospitals to offer digital healthcare services that were unaffordable to most; the cloud allows smart governance to leverage algorithms and data to answer problems no human can approach today; etc.
--- I am very disturbed by the lack of a comparison between the waste with and without the cloud. Sharing resources should intuitively lead to higher utilization than having everyone own their computers independently of each other, leading to efficiency and relatively less waste. (Growing need through affordability, akin to Jevons' effect in coal use, could counter this argument, but at the moment we do not know if it applies to computer use, and also the author does not bring this into discussion.)
--- The claims of proof made by the author ("we have shown", especially in the Conclusion) use scientific formulation, but for a trained scientist lack evidence beyond mere anecdote and circumstance. For me, they are devoid of value and a major negative part of the book.

[Mike B · Rating 3 out of 5]

What is a network?

Given how many things we use that term to describe, surprisingly little consensus exists about what it actually means. Connections, sure. But connections to what, and what kind of connections? With no clear-cut answers, interdisciplinary research has often been stunted by conflicts pitting syntax (i.e. neuroscience, archeology, genetics) against semantics (psychology; anthropology; epigenetics), in the battle for ideological supremacy.

Hu enters squarely on the semantic "B-team". For him, networks are central, but hardly the end all be all of modern media. Rather the important thing is how they are mediated, and become intelligible as an abstract but accurate political force.

Drawing on the "cloud" metaphor -- deployed initially by Bell Labs engineers, then given new life by Apple, Google and Amazon -- Hu argues for the existence of a "third space" mediating relationships between the virtual and material worlds. Key here is the cloud's ambiguous composition. Halfway between metaphor and artifact, it is both a material and immaterial place -- the shared home of all connected objects, on all Internets, ever.

If this seems peculiar, it actually has quite a bit of precedent. Think of what physicists call "the Universe": we don't really know what, or even if it is. All we really can say for sure is that it is where everything we know, and possibly ever will, seems to live. Theoretically speaking, this gives us the foundation on which to erect a systems map for global media: networks (core) sit inside of the cloud (periphery), each (re)producing and also (re)produced by the other.

How much this matters depends on your comfortability with essentialized abstractions. Early on, Hu recalls the time, while repairing network cables in early-90s Silicon Valley, he almost blinded himself by glancing into the open end of a live fibre-optic. The lesson is clear: stare too deeply into the network, and you will lose sight altogether ["That I can still see today is a testament to both to my dumb luck and, metaphorically, to the paradox that the cloud represents: that you can never see it by looking directly at it" (xx)].

But the argument here is essentially strategic. Ultimately, the cloud is still a way of interfacing with "the network" -- whose existence and singularity we simply take for granted*. Stepping back from the network, as an analytic, helps us see it as a historically constructed artifact. But it also creates an imperative to define what the network actually is. Are all networks, "the network"? And how can we prove it, without staring into the frayed ends of live cables?



* To his credit, Hu notes this problem and cautions against it -- but also fails to locate a way around it.

[Nick Ziegler · Rating 4 out of 5]

Despite some questionable and distracting theoretical deployment, to read this book is to begin cutting the first teeth in a key for understanding our present and future. I wish Hu were more forceful about underscoring the ways sovereignty is still operative -- indeed, I think his exploration of networked power and "control" pushes us to realize that "control" is always parasitic upon sovereignty, which problematizes the analytical distinction in the first place. (Even as I use power, I think this keyword really needs to be investigated whenever it's used, especially in the vague Foucaultian manner so common in the human sciences today).

The most useful concept I drew from this book was that of the "user" - a subject emergent from the experience of personal and cloud computing and which interfaces seamlessly with liberal ideologies of individual personhood, responsibility, and agency. This thematic of the cloud being coterminous with earlier cultural and technological forms is reiterated in an impressive number of contexts throughout the book. In this way Hu militates against easy narratives of progress and supercession -- a kind of "industry history" to rival the "Church history" of the past.

What isn't immediately emergent from the book is Hu's solution to this problem of how the "user" reifies and universalizes what was previously merely a hegemonic option for subjectivity -- that is, how to become the "public" he valorizes in the last pages of the book. If "public" is the palimpsest of "user," this needs further theoretical work. Why are the skywatchers seeking satellites or IRC hacker collectives like Anonymous not "publics," but still captured by the subject position of "user"? I don't doubt that this is the case, but what is the specificity of a "public"?

As much as Hu locates the discourses and infrastructure of the cloud outside the specific technologies that currently typify it, his answer seems to require something like a gathering in "meatspace." Hu suggests, for instance, that closing the internet in Egypt perhaps buttressed or "empowered" the collective subject called Tahrir Square. As Hu talks about leaving a paranoiac model of knowledge (which mimics the cloud surveillance apparatus) in favor of trusting "appearance," I'm put in mind of Ranciere's notion of politics in "Disagreement," in which it is the appearance of a subject not belonging to the order that constitutes the argument over the very conditions of discourse and material distribution. What kind of aleatary, unpoliceable apparition is still possible in a society of such metasticized control-sovereignty? And if we are all "users," can we even dream of an alternative?

[Benji · Rating 4 out of 5]

Digital crypts are less technologies of the future than ruins in disguise.

[Chris Esposo · Rating 4 out of 5]

This is a very different take on the history of “the cloud”. I was initially expecting a literal history of the technology that would eventually become the cloud, similar to a history of the first internets as recounted in the book “Where Wizards Stay Up Late”. Instead, this book is an analysis of the “concept” of the cloud, and here there’s a bit of inversion with respect to topic, as most of the public understands the cloud as a relatively new concept, an infrastructure that supports the era of Web 2.0 and ‘big data’.

However, the author’s contention is that the cloud is a very old concept, and what we refer to as “the cloud” is merely the current material implementation of this old concept. As described by the author, the ‘concept’ of the cloud is a combination of two different concepts, one being the notion of the “network” with less emphasis on material connections between nodes, but just that there are nodes at all, and the second concept, the other being a bit ‘fuzzier’ notion of disintermediating privacy through anonymity via a network structure.

The author uses the pre-digital, industrial era-innovation of the urban plumbing system as an early example of an “cloud”, since it was both a network, and the usage of it provided a kind of anonymity to it’s users, since the refuse was aggregated eventually to one flow by the system. In a similar way, the VM-driven cloud computing architecture also satisfies these two criterions laid out by the author, making it possible for two people to not even know thy are calling software housed on the same disk (or leveraging the same memory as a competitor say).

This sort of analysis is novel to the field of computing, and is more akin to “critical analysis” as found in literature or history. As computing has penetrated domains outside of it’s own strict subject matter in the past decade however, an analysis focused on the societal context and/or the materiality of computing objects have become more popular fields of studies in academia (and some aspects of industry), another similar book in style to this one is Ed Finn’s “What Algorithms Want”, which was also an excellent read.

In assessing the concept of the cloud this way, the technical reader is able to better make apparently disparate connections between their craft and other subject-matters, as well as better understand the ramifications of the objects they are implementing. This ability to make context will become more important, especially as computing permeates “the edge” via IoT and other physical devices/agents. In a way, the analysis outlined here is the first step to developing an ethics for computing, which will again be critical in the age of AI.

Overall, there should be more books like these, and I hope that this and other authors continue to write/think about modern computing objects in this way. Recommended

[laurent · Rating 5 out of 5]

absolutely fucking incredible. the author being a chinese network engineer turned poet and media scholar living in europe - my god he justtt like me!!! first chapter analysing the political symbolism of network diagrams got me whooping and cheering in my room like i was watching football. i really love the specific technical language used - things i remember from computer network class - and i realised it wasn't that i was bad at understanding networking technology, it just felt like mythologies way too good to be true... recognising a pattern in my life where i simply cannot follow instructions i don't understand... my brain and body refuses to cooperate no matter how hard i try... to me quantum physics is realer than o level physics and for the life of me i couldn't do o level physics but at 16 i understood quantum physics theorem really quickly shit was just way easier to comprehend...

[Peter · Rating 3 out of 5]

On one hand, I really enjoyed much of what this book of theory evokes in terms of building onto the infrastructure of the past. However, I really struggled to understand how the author’s examples played into the broader idea of data sovereignty. I think when you start dropping Deleuze into your thinking you gain a poetic broadness while sacrificing clarity. Maybe I’m just getting dumber as I get older, but I appreciate a well-articulated idea.

I originally picked up this book when I was working as a tech recruiter, and back then (8 years ago) I was far more forgiving of abstract, artistic theoretical conjecture. I am less so now. Read this when you’re young and just exiting the Obama era!

[Ben · Rating 3 out of 5]

I saw this as a good albeit more abstruse and eclectic (part of the appeal IMO) complement to Tim Wu’s The Master Switch on centralizing forces underlying the internet. Where Wu is focused more on antitrust and economic entities, Fu is focused more on political bodies of power coopting the internet to atomize and encroach upon us.

[Caroline · Rating 5 out of 5]

This was phenomenal

[spoon · Rating 4 out of 5]

Emphasis on materiality is 10s 10s 10s across the board. Also loved emphasis on cloud as built on fear of loss. (Got to page 111 of 148)

[Mynt Marsellus · Rating 4 out of 5]

The middle chapters are really weak but this is a must read for contemporary computer users (aka all of us).

[Lucy · Rating 3 out of 5]

Well, for a book titled "A Prehistory of the Cloud," this is not the book I thought it would be. I thought Hu would offer a history of how the Cloud came into being -- from phone lines to the early DoD Internets, to the collecting and selling of user data, to the dominance of AWS (Amazon Web Services).

Instead, Prehistory reads mostly like a book of critical theory, interested mostly in interrogating the semantics of certain phrases like "sovereignty" and "networks," and close-reading semiotics such as the "cloud-like" shape of engineering diagrams that depict network infrastructure.

In no particular order, a list of what I considered to be the key takeaways.

1. There is no singular Internet: there are multiple Internets.

In fact, in 2010, then US Secretary of State Hillary Clinton declared both Internet freedom and a singular, global Internet as "pillars" of America's foreign policy. In fact, the US holds the leadership position in the Internet oversight organization ICANN.

On the simplest level, the Internet accessible in mainland China is drastically different than the Internet accessible in the US.

Slightly more complicated, we tend to think of area Internet shutdowns as authoritarian tactics used by foreign states: the Egyptian state, during mass protests in the winter of 2011, cut cell phone service for a day and Internet service for 6 days; local governments in India regularly cut cell phone and Internet service to "try" to abate Whatsapp-fueled storms of misinformation and pitched fervor. But it happens domestically in the US as well: in 2011 after a BART (Bay Area public transportation) police officer fatally shot a homeless man, BART officials preemptively shut down cell phone service system-wide (airports, commuter rail, subways) to successfully prevent a mass protest.

2. The cloud creates a sense of computing power as a virtually unlimited resource, but really, the cloud masks hardware with software.

For those who don't know, "the cloud" is a huge number of remote computers that sit in rural areas, and consume an enormous amount of water and fossil fuels. Your childhood photos are not zipping around formlessly in some electromagnetic ether; instead, the world's data centers collectively are 2% of the world's electricity and 3% of the world's carbon footprint--on par with the entire aviation industry.

3. The cloud is a neoliberal fantasy about user participation.

The "new" shifts to tracking user data and charging for computing time (a la AWS) are better understood as a reversion to standard practices in early computer/Internet days. In the 1980s, multiple computer users would run programs on the same computer, and they would be billed by individual usage. The user anonymity and "one-time lump sum cost" of 90s and 2000s PC and Internet culture are the real anomaly.

4. Everything is big data now: consumerism, journalism, mass surveillance, destabilizing societies, and even war.

[Casey Browne · Rating 3 out of 5]

The basic premise is that the metaphors through which we understand today's internet are actually built on metaphors and infrastructure that are much older. These metaphors and infrastructure structure how we think about the internet, naturalizing particular ideological constructs of the internet.

I thought that this book made some interesting points. Still, the fact that he is conflating infrastructure and discourse (that is, metaphor) without really teasing out why these two very different things should be juxtaposed, seemed deeply troubling.

For much of the book, I was also wondering, "So What?" What does it matter if the internet was built along railroad tracks? What does it matter if there was this idea for a series of vans travelling around the US, downloading and uploading videos to get around television's centralization?

At some points, he seems to suggest the importance of his project. He says that the sewage system interpolated a kind of modern subject position that was a combination of public and private, which helped create the contemporary subject position. But, he does not really seem to explain why this nexus between older metaphors or older infrastructure is important for much of the time.

[Eve · Rating 3 out of 5]

I wanted to love it but somehow I just couldn't.

The arguments (mostly) felt a bit old hat but there was a factoid or two that grabbed my attention.

[Anicca Cox]

Mind blowing.