Lightbulbs and ducks: The power of tech platforms to make waste disappear

Data, data, everywhere

Every time we use our computers to search for information, find a restaurant, play music, download a video, watch a YouTube clip or watch Netflix, we both access and create data. Every keystroke, every request, text, personal message, photograph, email and our every movement are recorded. This is the era of Big Data in which data is collected and processed in high volume at velocity. ⁹ But this obsession with data collection, use and storage has a cost. Google estimates that a single search using its services produces approximately 0.2 g of carbon. ¹⁰ Two hundred milligrams is a breath, a puff of carbon. However, when you combine this information with other available information and do the mathematics, a problem appears. Google processes ‘more than 40,000 searches every second (3.5 billion searches per day)!’ ¹¹ That works out to be 700,000 kg or 700 tonnes of carbon released into the atmosphere every day just on Google searches. But search is only a small part of the problem. International technology company, Cisco estimated that ‘video will make up 82 per cent of internet traffic by 2021, up from 73 per cent in 2016. Around a third of internet traffic in North America is already dedicated to streaming Netflix services alone’. ¹² Cisco also reported that, in 2017, ‘global [internet] traffic was 122 EB [exabytes] (over one billion Gigabytes [GB]) per month’. ¹³ That number is expected to triple by 2022 so that global internet traffic will soon be 396 EB per month. ¹⁴ EMC, an international data corporation estimated that in 2016 ‘more than 1.8 trillion gigabytes of digital information were created globally’. ¹⁵ Each day we create 2.5 quintillion bytes of data and this is set to multiply as the internet of things (IoT) expands its reach. ¹⁶ Ninety per cent of the world’s data was created in the last two years. ¹⁷ And the pace is not slowing as we create more sophisticated, magical and elegant machines – and our appetite for novelty wrapped in beauty increases.

The number and variety of points from which data can be collected also continues to grow. ¹⁸ Autonomous vehicles will generate 4 TB of data in about an hour and a half of use. ¹⁹ However, that data is not static. The computers in the automated vehicle will have to constantly interrogate vast amounts of data, data processed through data centres, in real time. One Intel director described the ‘in-vehicle compute as, “rather like a second engine within our car”’. ²⁰ Since the demise of the use of film in cameras, anyone with a phone can take as many photographs as they like, and often of the same thing. Infotrends estimated that in 2017, people took 1.2 trillion digital photos. ²¹ These digital images are syphoned off to be stored in the cloud. Now, we converse in Tik Tok videos, memes, emojis and video clips, each requiring more space to store and more power to process. In an effort that epitomises platform entity quests for power, Google sought to digitise every book on the planet. ²²

The size of the data black hole that we have created is incomprehensible to us. It is lost in a miasma of mathematical shorthand; what is a quintillion? It is 10¹⁸ of course – but it is also 1,000,000,000,000,000,000 and we create 2.5 quintillion bytes of information every day. What can we do with this much data? How can we store it and how can we process it? Should all data be stored? Is all data equal? Part of the problem is duplication. Data is not just stored once. There are multiple copies stored in multiple places to ensure the data is never lost. I have copies of my data on multiple storage devices. Should my wedding photos have an equal right to exist in these formats as a digitised version of the King James Bible, or the Magna Carta for example?

In the subheading for this part of the article, the allusion to the Rime of the Ancient Mariner – and thus to destroying nature – is apt. The mariner was only saved after he regained his respect for nature. We find ourselves in an analogous situation; lost in a sea of data, most of which we do not need, or cannot use, having rejected nature, an albatross firmly around our necks, and nature against us in the form of vast arrays of data centres blazing against our humanity. This article is not a warning about rejecting nature (but it could easily become one). Instead, it focusses on the means by which we are distracted or distanced from the impact or our elegant online lives and how we became distracted from the mischief inherent in our digital lives. It asks how the beauty of the products we buy distances us from the waste we create. That waste manifests in places other than our homes. We are removed from it because the data we consume creates waste, but not in our lounge rooms. I discuss this further later in the article, but first, what do we do with all that data?

Data centres – our dirty secret

As the large technology platform entities and their appetite for data has grown over the last two decades, so too has the need for more and more data storage and processing equipment. As the data we create has grown, so too has our need for storage space and in the 2010s, businesses needed to create data storage rooms. Apple stores immense (I have run out of superlatives) volumes of data, although it calls its storage facility iCloud. When its customers sign up for iCloud, they get 5 GB free space and can then buy more if needed. Amazon captivates us with bargains and unlimited access to everything. Yet, its systems rely on, and create vast amounts of data that has to be preserved in the cloud. Amazon offers personal and business cloud storage services. Amazon Web Services or AWS is now the largest and most used cloud computing facility in the world. ²³ Facebook employs psychiatrists to ensure that we are constantly using its product. Its streams of information and video require data to be at the ready for every subscriber at any time. It also stores vast amounts of user data in the form of photos and videos and in this way is effectively a cloud storage company. ²⁴ Many data warehouses are in remote locations, out of sight of the end users.

As the data has multiplied, like the brooms of the sorcerer’s apprentice, the platform entities began building warehouses to house the ever-multiplying storage and processing hardware. Just as personal computers emit heat, these rows of stacks of servers and storage devices create enormous amounts of heat. However, heat waste created in far off warehouses does not raise the same concerns for us that physical household or construction waste does because it is invisible; it happens somewhere else; and it does not require a space (such as a hole in the ground) to store it. But it is real. Like some paradoxical mythical beast, these warehouses filled with stacks of servers have become monsters whose purpose (like the incandescent lightbulb) seems to be to create heat, but that need to be cooled lest they burn out. Most of the time, these servers are not processing data. It is an irony that it is the systems used to cool the warehouses that use the bulk of the energy that data centres consume. Because we demand that information must always be available at the touch of the screen, the servers must always be running, ready to spit out the data when required. Glanz noted that typically, the energy wasted in having the servers constantly ready to process data ‘is as much as 30 times the amount of electricity used to carry out the basic purpose of the data center’. ²⁵ One study estimated that only six to 12 per cent of the power that data centres use is for computation. ²⁶ The purpose of the process is to store and process data, but the major output is wasted heat. Again, the comparison to the lightbulb is apt.

Like the facts about the data itself, we must resort to comparisons with things of which we at least have some understanding because a straight recounting of the facts would bewilder us. Some examples of this tendency to compare include that: ‘a single data center can take more power than a medium-size town’; ²⁷ Chinese data centres ‘consume more electricity than all of Hungary and Greece combined’ ²⁸ and; data centres in the United States (US) alone now consume more than ‘90 billion kilowatt-hours of electricity a year, requiring roughly 34 giant (500-megawatt) coal-powered plants’. ²⁹ This consumption is expected to double every four years. ³⁰

By 2025, the world data centre industry is set to consume approximately 3000 TWh of power per year. That seems like a lot but do these comparisons help us to comprehend the size of the problem or to conceal it in a blur of bigness, like the ‘Galaxy Song’ in Monty Python’s The Meaning of Life. ³¹ This use of the mathematician’s shorthand elides the true scale of the energy consumption – until we are told that this equates to the current electricity usage of the US. ³²

Because of our constant need for instant information and the impact that any delay might have on our economy and our lives, these data centres must never be allowed to go offline, even for a short interval. The data centres are therefore supported by rows of diesel generators that idle along next to the data centre just in case there is an outage or a blip in the electricity supply. The New York Times reported that Amazon had applied for permits to run enough generators at its data centres in a small county in Virginia in the US ‘to nearly match the output of a nuclear power plant’ ³³ – and that is just the backup power. There are over three million data centres across the globe. ³⁴ Again, to allow us to comprehend the enormity of the issue we must resort to analogy. But this article is not about the use of metaphor to comprehend the incomprehensible (although, again, it easily could be). Instead, it seeks to understand how we as consumers have come to accept, or at least not reject, this gross inefficiency and waste.

Waste as a by-product of modernity

Hawkins framed her arguments on the ethics of waste around disposability, distance, and denial as the three enablers of waste. ³⁵ She argued that we have developed habits, or an ethos of disposability brought on by the growth of industrialisation. She noted that the increasing overabundance of commodities ‘marketed as convenient and disposable evoked a modernist asceticism and temporality in which the technical was valorized for saving time and for its instrumental rationality’. ³⁶ To be able to morally accept the levels of personal waste, we have distanced ourselves from our waste; it is ‘taken away’. ³⁷ Distance is the physical space that we place between us and our waste. From the moment it becomes waste, ‘we are distanced from the material substance of things’. ³⁸ The tangible becomes intangible. Hawkins says that ‘while we know generally where it goes, the invisibility of these places, their location underground or on the margins of cities, facilitates denial or active not knowing’. ³⁹ Hawkins was talking about physical waste, but her analysis of the human relationship with waste also applies, perhaps doubly, to waste caused by data collection, use and storage; especially since we cannot see the waste these processes cause; there is no physical trace from which we must distance ourselves. Hawkins expands on distance:

Consumer cultures and the technocratic logics of efficiency and concealment have produced a distanced relation with wasted things even as amounts of waste have escalated phenomenally. External systems of removal from garbage trucks to sewers have dramatically reduced the demands waste makes on us. ⁴⁰

Again, Hawkins was talking about physical waste or garbage, but heat waste is invisible, intangible and therefore all the more removed from our senses. The distance is compounded because the products we use are elegant, almost magical. We are mesmerised by their beauty that further distances us from any waste. Modern technology products are designed and marketed to be attractive to our senses. New products look, feel and even smell good. This desire for technological cachet drives people to sleep outside Apple stores to line up to buy the latest version of the iPhone. The elegantly designed products of technology companies distract us from the waste that supports their effectiveness. To most consumers, there is no waste. These products are pure efficiency. Simply plug them in and begin accessing the world of data that waits at our fingertips. Their sleek and beautiful designs distance us from causation. It is invisible – like the infrared light in the form of heat from the lightbulb.

How do you solve a problem like data centres? Control the discourse

On closer inspection, the many and varied techniques used by platform entities to accumulate and enforce their power become clearer. They include controlling the discourse surrounding the technology, using dark patterns to coerce behaviour, and creating an imaginary in which the elegance, attractiveness, and ironically, the power of the machines and the platforms ⁴¹ are indomitable, insuperable and indispensable artefacts. It is in this imaginary that platforms can exude power and control politics.

In the last decade or so, we have been coerced or directed to store data in the cloud. What is the cloud? The cloud is so-called because in schematic drawings of computer systems, the internet is often depicted as a cloud. However, the use of the term has proved to be effective to create the illusion of environmental propriety. The cloud sounds innocuous, ethereal; even natural. Clouds bring rain that is essential for life. A cloud is delicate, cool and cooling – at dew point. Gods exist in clouds. Clouds are intangible; they have no physicality. Of course, we want to store our data in a cloud. We have lined up to hand our data to the cloud. However, cloud storage or cloud computing is not all these things.

As mentioned, the ‘cloud’ is really a conglomeration of overheated computer hardware. It relies on vast banks of computer storage space set up in massive warehouses in disparate parts of the world. This semantic sleight of hand is more than a little misleading. Users of technology are beguiled into using more and more of the technology that causes the heat waste. By naming this phalanx of data centres the ‘cloud’, the giant platform entities – Google, Amazon, Facebook, Apple and Microsoft – have captured the discourse to create a more serene picture than exists. In this same vein, data warehouses are sometimes referred to as data farms. As with the ‘cloud’, the irony is staggering. Farms, like the cloud, are synonymous with healthy lives, open spaces, freedom to roam. However, this is not the allusion that is intended. The reason that ‘data farms’ is used to describe data centres is because of the banks of computers lined up in rows like battery farm hens and feed lots that have replaced open farming. This blight completes our rejection of nature and requires us to reject reality to accept the elegant products and unlimited service that the technology companies offer. This requires us to deny the reality of waste – something with which the technology companies are all too happy to assist us.

None of this nomenclature is created by the platform entities, but it is certainly used as part of the discourse that promotes access to unlimited information and power through small clean designer boxes. They do not clarify or reject the connotations associated with clouds and farms and neither do they openly discuss their dirty waste secrets. None of the statistics in this article come from the platform entities. Instead, they come from investigations into heat waste by diligent researchers and reporters. This leads us to consider the power dynamics between the platform entities and users. Studies of power also raise questions about the role of law in creating the atmosphere in which power flourishes and law’s role in creating and protecting power.

Power and law

Law is less adept at addressing exercises of power, especially when they are not overt or do not immediately cause damage. Lukes’ third dimension of power considers the way that power often operates outside the view of observers; it may not even be exercised at all. ⁴² He argued that this dimension of power shapes people’s preferences, cognition or perceptions without direct conflict or overt control or manipulation so that the less powerful ‘accept their role in the existing order of things’. ⁴³ This raises proof and causation problems in law that foreclose immediate legal solutions. Addressing this, Gaventa provided indicia of third dimension power in situations of inequality such as through using social myths, the social construction of meanings and patterns, adaptations to powerlessness like learned helplessness or fatalism, or being socialised into compliance. ⁴⁴ So identifying these types of situations in the interactions between platform entities and their users may reveal characteristics of the exercise of third dimension power. I have elsewhere provided examples of the use of third dimension power techniques by the platform entities, including controlling the discourse around the technology, and argued that these are normalised over time. ⁴⁵ In exercises of third dimension power the powerless either do not experience the power exercised over them or feel helpless to alter it if they could perceive it.

The Collingridge dilemma on the control of technology states that, at the earliest stages of development of a new technology, regulation is difficult due to a lack of information, while in the later stages the technology is so entrenched in our daily lives that there is a resistance to regulatory change from users, developers and investors. ⁴⁶ Thus, we appear to be stuck with data centres. They have become part of the fabric of our technological existence and we cannot do without them. The platform entities seem to have learned from the Collingridge entrenchment playbook ⁴⁷ – not as a warning to take things slowly, but to rush new products to market, to ensure that their products are entrenched before any objections can be raised. In this way, technology companies are not in breach of any law. They remain within the confines of extant laws. But misuses of power, the mischief of the data centres, while technically legal, must be exposed and redressed. Regulatory responses to this mischief can have some ameliorating effect.

Black defines regulation as ‘organised attempts to manage risks or behaviour to address a collective problem or concern’. ⁴⁸ It must be organised and not simply organic or accidental. This definition includes laws, but it is not limited to purely legislative responses. Lawrence Lessig argued that there were four modalities of regulation: law, norms, the market and architecture. ⁴⁹ This characterisation provides a good framework to address regulatory responses that include the use of law.

I will now address Lessig’s four modalities in turn. Governments resist intervening into private ordering. Businesses are generally left to run the way they see fit without state intervention if they are not in breach of any laws. Platform entities will attempt, for as long as possible, to assuage concerns with notions of self-regulation. However, Braithwaite acknowledges that there is ‘formidable history of industry abuse’ of the privilege of self-regulation without the threat imposed by the ‘benign big gun’ of government legislative intervention. ⁵⁰ The heat waste associated with platform entity operations will ultimately require a legal solution. In terms of laws, environmental laws govern the polluters, not the wasteful over-users. The power generators must comply with environmental controls that regulate pollution levels, but the platform entities are removed from this regulatory constraint and can use as much power as they require as long as they pay for it. So environmental laws could be extended to large users to ensure the most efficient use of power or to mandate renewable energy targets. Lessig proposed that the law could be involved to facilitate each regulatory modality.

Regulation by market operation is another of the modalities that Lessig recognised. The price of electricity itself does not seem to be a sufficient deterrent as the use of electricity and the number of warehouses continues to rise. The platform entities simply factor this cost into the price to advertisers using their services and their profits continue to grow year on year. Neither environmental laws nor the price structure is sufficient to result in a change of usage and wastage. But an examination of the treatment of externalities in competition law illustrates that they have long been a trigger for market intervention – that is, governments intervene to correct the damage caused by the externality. This article suggests that such a regulatory intervention into this market to perhaps increase the price of largescale waste of electricity is required to reduce the waste caused by modern platform entities.

The power of the platform entities used to conceal this environmental problem requires a collective normative regulatory response that will include education campaigns that lift the veil of innocence between the users of technology and the waste it creates, and some proactive solutions from the technology companies themselves. The costs of heat waste as pollution are borne by society, while the platform entities avoid accountability. This is why platform entities must keep the veil in place. First, to continue to promote the understanding of technology as only beneficial, but also to remove the gaze from the very real damage caused by their operations. However, normative responses to pollution are shifting so things that cause waste are becoming more apparent and less acceptable to the majority of the population. One suspects that the threat of user understanding and reaction is partly why the platform entities have steered the discourse away from the damage it causes and highlighted the magical, techno-utopian aspects of its products and services. In terms of laws reinforcing normative order, laws could be developed to force platform entities to publish their electricity usage which would allow users to decide to punish or support more efficient electricity users. This would also incentivise the platform entities to innovate and develop more efficient energy use in the architecture of their systems.

In terms of architecture, the platforms are seeking ways to reuse or reduce their heat waste through various mechanical means. Reducing the amount of power that data centres use, and therefore the amount of heat they generate, is one way to attack the problem. On 20 July 2016, concerned by its energy waste, Google reported that it had deployed DeepMind’s machine learning in a series of tests on one of its live data centres. ⁵¹ The tests resulted in a reported 40 per cent decrease in energy consumption for the centre while the AI was applied. ⁵² Another possible partial solution is to find a use for the excess heat. There are proposals in both Stockholm ⁵³ and Amsterdam ⁵⁴ to funnel excess heat from data centres into the cities’ underground heating grid which could ultimately provide 10 per cent of these cities’ heating. ⁵⁵ Further, the use of renewable sources of electricity such as solar and wind power are being explored, to attempt to reduce the carbon footprint of data centres. Again, this could be mandated by law if sufficient progress is not made.

Conclusion

Data centres create vast amounts of heat waste and must be cooled. Unless we can change our obsession with Big Data or find more efficient methods of storing and retrieving it – like the LED light that has replaced the incandescent bulb – our digital lives will also continue to create outsized waste as heat and CO₂. Solutions might take the form of regulation to ensure that the big technology companies do not just pay for the electricity service, but also reduce their energy consumption for broader societal interests. It will take engagement and an informed commitment from regulators and users to break the beguiling spell of beauty and wonder that has so transfixed us and lulled us into our collective wasteful ways. The power of the big technology companies to control the ways we perceive their products, and their by-products, requires collective action and perhaps an acknowledgment that with great power comes great responsibility.