What Tung-Hui Hu’s “A Prehistory of the Cloud Teaches us About Cloud Marketplaces
As a former network engineer and current professor of cultural studies, Tong-Hui Hu offers readers a unique approach to understanding the cloud in his book, A Prehistory of the Cloud (2015). Drawing from philosophy, history, and documentation from the hyperscalers themselves, Hu’s book is dense. Yet he offers some of the best definitions and descriptions for terms we throw around in the cloud marketplace world, from the “cloud” itself to topics like consumption, Google search, data centers, and virtualization.
Here’s a look at some definitions and descriptions of key terms in the current tech landscape that can be useful for technical and non-technical audiences.
What is the cloud?
Where does the cloud exist? In Hu’s book this is a big topic of discussion. We often imagine the cloud as something in the air – “inexhaustible, limitless, invisible” (61). In reality, the cloud is created by a robust network of servers in data centers, many of which are located in rural areas and resemble large factories.
When we’re thinking about the cloud created by hyperscalers like AWS, Azure, and Google, Hu encourages us to imagine the cloud as “a city of gated communities.” So much of the world is connected by a singular internet, yet we have access to a limited number of communities.
With the power of the cloud, users are able “to call forth as much compute power as one needs from a vast set of central servers that have access to virtually unlimited computer resources” (66). The cloud, then, effectively “masks hardware with software” (146). Cloud storage, as well as cloud-based applications like the ones available on cloud marketplaces, “smooth out the chaos of the internets and produce a singular Internet.”
Data centers
If data centers were their own country, they would be the fifth largest energy producer in the world – more than Canada or Japan!
Hu describes data centers as “miniature cities of computer servers and hard disks that enable data to be stored and concentrated ‘in the cloud’” (59). For those of us that remember the days of bulky computer monitors and large CPU towers, data centers effectively replace these technologies. With cloud computing, users scattered across the region rely on the same hardware to complete tasks
Originally located in urban centers, data centers have since migrated to rural and exurban areas where land and utilities are cheaper, Hu notes. That’s why Meta, another hyperscaler, has plans to build a new data center in Louisiana.
Google Search
As an SEO nerd, I am always looking for new ways to describe how search works. In Hu’s book, he describes how Google’s search algorithm can offer different results when the same keyword is searched at the exact same time by two different users:
Search Google two times for the exact same keywords and you might expect the same results. But if you are logged in as one user, the word ‘nature’ might return results related to fishing and hunting, while a second user might receive results related to environmental justice. Algorithmic filters cause search engines to return different research results, depending on your imputed gender, demographic, and social class.” 65
This is only becoming more relevant as users conduct more searches on LLMs like ChatGPT, Claude, and Perpexity, which offer more customized but less predictable results.
Consumption
When crafting a “better together” story for co-selling, explaining how your cloud marketplace solution drives compute consumption can help you get attention from sales reps at AWS, Azure, Google Cloud, and Salesforce.
In Hu’s book, he explains that, originally, consumption and “computing costs were generally bundled into the price of a new computer or software package” (40). This approach, of course, had its limits.
When cloud applications hit the scene, business leaders and technologies began imagining a world where consumption could be almost infinite. As Hu explains, “Regardless of how much bandwidth costs, and how much actual money is spent, the underlying logic of freeway capitalism is consumption–of time” (51).
While consumption on cloud marketplaces is slightly more complicated than time alone, the main point holds: the more time users spend on your application leads to more tasks being completed, more compute being processed, and more revenue for cloud service providers (CSPs) like AWS, Azure, and Google Cloud.
Virtualization
If you spend some time poking around AWS, Azure, or Google Cloud documentation, you’ll probably encounter the term virtualization.
To describe this, Hu builds on his metaphor of the cloud as “gated communities.” Virtualization allows our private virtual worlds to become one global city. As Hu describes it, “Virtualization is a way of constructing a simulated environment that both allows a user unprecedented freedom (it seems as if she has control over an entire virtual environment) and restricts that user from ‘leaking’ or contaminating the data of other users.” (61) In other words, virtualization allows millions of users to share a data center in the cloud.
Hu pushes us to imagine today’s cloud infrastructure as a Manhattan skyscraper – different people from different businesses and backgrounds might cross paths in the lobby or elevator, but spend the majority of their time in private environments, with limited access. “Similarly, inside a data center, data are connected only paratactically next to each other…military documents next to banking records next to your email” (71).
Thanks to virtualization, it’s “impossible to sense the other users on the same storage device, server, or network.” Each user has the impression that they have their “own server on the cloud that can be instantly created on demand” (62).
Now imagine what it would be like there was a fire drill in that Manhattan skyscraper, and workers that would never cross paths during the work day pour onto the street and begin mingling together?
Virtual Machines
Thanks to virtualization, tens of thousands of clients can host “virtual machines” – e.g., their own instance of an application – on the same physical machine. Virtualization software ensures each virtual machine is isolated from the others, ensuring privacy for clients.
IP addresses
Hu suggests that we think of IP addresses as “something like a phone number for your computer.” And just like how certain zipcodes have run out (like 212 in Manhattan), the boom in personal computing means that the old IP address format ran out of options. Today, new IP addresses are longer, called IPv6.
Data Backup and Recovery
Before the cloud, the first Data Backup and Recovery companies stored data as hardware in a vault, in a mountain located in upstate New York, “designed to survive war or an atomic explosion” (96).
Today, solutions like Veeam, AvePoint and others leverage the cloud to safeguard their business’s environments. As Hu explains, Disaster Backup and Recovery solutions “can mirror or replicate a company’s entire computing environment – often just not files, but also entire operating systems – to the cloud; when a disaster strikes, a virtual machine in the cloud simulates a failed computer’s setup. If the virtual machine does this fast enough, clients may not notice the failure at all; the server is said to ‘fail over to the cloud’” (96).
Data Backup and Recovery solutions also allow users to “back up and archive exabytes of infrequently used data, data that does not need to be ‘live,’ or immediately accessible” (96).
Drive Cloud Marketplace Success with Invisory
Interested in chatting through your cloud marketplace strategy? Looking to get listed or drive co-sell success on AWS, Azure, Google Cloud Marketplace or the Salesforce AppExchange. Let’s talk!
