Enabling the 98%: The role of Sustainable IT in the modern Computer Science syllabus

Abstract

Sustainability is a theme of increasing global prominence. The ICT sector is responding in two important ways. The first is to reduce its own carbon footprint, estimated at 2% of global emissions. The second is to concentrate on the rest of the economy (the 98%) and to produce innovative products and services that have a positive sustainability impact. In this position paper I show that various sustainability perspectives allow us to embed the paradigm across the whole computer science curriculum, and to use this work to deepen links with other disciplines. We should avoid getting drawn into climate science debates that will detract from the value Sustainable IT can add to our students’ education. Nevertheless, sustainability offers an opportunity for us to engage the next generation of technologists and policy makers and equip them with the skills they will need to compete in the low carbon economy.

Keywords:

• Sustainability
• ICT
• Low Carbon Economy
• IT education

1. Introduction

The subject of sustainability has a general resonance at the moment, with political (United Nations, 2009), cultural (Hopkins, 2008) and industry (Gartner, 2009) interest at a seemingly all time high. With respect to the IT sector, we are, on the one hand, confronted with the size of our carbon footprint, apparently running at 2% of global emissions (Gartner, 2007), a figure which has erroneously been equated to that of airlines (Malmodin et al 2010). On the other hand, we also see a great deal of investment and commercial activity in the Green IT sector, which has recently seen a plethora of startups and serious investment by deep-pocketed players such as Tom Siebel (Ricketts, 2009).

How, then, should we in the IT education space respond to this new trend? In this position paper, I offer a possible definition for Sustainable IT that will naturally lead to some potential approaches for teaching the subject. Its importance can be underlined by empirical industry trends, which can be usefully subdivided into technologies that (i) reduce the impact of the ICT sector (the 2%), and (ii) have a positive sustainability advantage on the rest of the economy (the 98%). There are many areas for fruitful discussion within the IT curriculum, although there is also a danger of getting derailed by climate science debates. Sustainability has the potential to build links with other subjects, and to act as an over-arching teaching and research theme. Perhaps most importantly, sustainability offers an opportunity for us to engage the next generation of technologists and policy makers and equip them with the skills they will need to compete in the low carbon economy.

2. What is sustainable IT?

The easiest way to define Sustainable IT is to consider the terms separately. “IT” stands for information technology, a term that is often conflated with “ICT” or information & communications technology. The difference of course is that the latter explicitly calls out technologies such as mobile phones and the internet. Of course, one could also cast the net wider, including entertainment technology such as games consoles and digital TVs. There is no single right answer (thus providing an immediate discussion opportunity); however most industry analysts tend to the middle, “ICT”, definition.

Sustainability is a notion well captured by the Brundtland commission definition (World Commission on Environment and Development, 1987):

“Sustainable development is development that meets the needs of the present without compromising the ability of future generations to meet their own needs.”

Thus, the challenge for our generation is to work out a way of living that delivers on our aspirations without comprising the quality of life of our children and grandchildren. This is a notion that can be delivered as an ethical proposition, but it also is one that has commercial and political resonance and hence a pragmatic significance. Moreover, the definition can be linked to philosophical and economic difficulties with defining and measuring well being, an area with a rich history of debate by thinkers like Bentham (Bentham, 1789), Rawls (Rawls, 1971) and Sen (Sen, 1987).

Sustainable IT, then, is the notion of an IT industry that has a reduced impact on — and even a positive contribution to — the sustainability of human civilization.

3. Why is Sustainable ICT important

The importance of sustainability for the ICT industry can be tackled in one of two ways. The first is an appeal to ethics, which is interesting and has much potential for a lively philosophical discussion. A perhaps more fruitful approach is to look at empirical trends in the industry.

Sustainability is both a constraint and an opportunity for the ICT industry. It is a constraint because of physical, legal and social requirements. Firms need to ensure that they are meeting legislative requirements for standards such as EnergyStar¹ or Blue Angel. Large companies must report (and reduce) their energy use through mandatory programs like the EU Emissions Trading Scheme² and the UK CRC Energy Efficiency Scheme³. There are significant reputational and financial risks for non-compliance (for the CRC, penalties can rise to a significant fraction of a company’s energy bill, and in extremis, a custodial sentence for a company director). In addition, organizations like Greenpeace are constantly monitoring companies’ environmental performance, publishing league tables and identifying laggards. Issues like the UK energy gap, peak oil and water shortages provide constraints that are at least partly down to the immutable laws of physics.

But sustainability is also an opportunity, especially for the IT industry, because of the potential to create innovative products and services aimed at the sustainability market. Some examples would be: travel replacement technologies (video conferencing, remote access and so on), eServices, paper reduction and virtualization. Smart Grid is a particularly significant trend. The term refers to a collection of technologies, from metering to control of supply, dynamic tariffs, energy storage and demand prediction. With the much-trailed UK energy gap fast approaching (Deloitte, 2006), the smart grid provides an important backdrop to future IT products and services.

Another significant theme is that of a shift to services, where individuals and companies use ICT to outsource and streamline their business operations. This trend is well documented in the analyst literature and has had a notable impact in the multi billion dollar Green IT services space (Mines, 2008). To take one example, the enterprise carbon dashboard market is a crowded space with many entrants attracting significant venture capital funding (Verdantix, 2009). There is a very useful video from WWF and Ashridge Business School (WWF, 2009) that would make a nice introduction to this topic.

4. The 2% and the 98%

I would now like to introduce a concept that I believe could act as a focal point for the teaching of sustainable IT. I started this paper by noting that the ICT industry has been charged with causing carbon emissions that amount to 2% of the world’s total (Gartner, 2007). This statement alone is an excellent starting point to explore various issues. Firstly, the present, historical and future scope of the ICT industry is a key concept that cuts across the whole syllabus in computer science. The boundaries are not always clear; do we, for example, include consumer devices (programmable fridges), infrastructure (smart grid), entertainment (games consoles), industry (robots) and so on? Another important point is that of lifecycle. From an environmental standpoint, the boundaries of ICT can be drawn around the manufacturer (called “Cradle to Gate”) or can be drawn further to encompass the consumer (the use phase) and of course end of life (recycling or disposal).

Bringing the discussion back to the 2% figure, Gartner adopted the ICT definition outlined above, including mobile phones but excluding games consoles. In addition, the analysis is necessarily time bound; the expectation is that data centre growth will push the total closer to 3% in the next few years. With regards to lifecycle, they based their analysis on the entire lifecycle of the ICT industry — albeit with a number of (necessary) assumptions and estimates. Malmodin et al (2010) explain why the comparison with airlines is distorted, since the latter analysis only takes into account fuel use and does not consider production and end of life, let alone the (controversial) issue of contrails. I would advocate resisting any temptation to use this analysis as a stick with which to beat the airline industry. Instead, I suggest it is a valuable illustration of how important context is when interpreting data. It also acts as a cautionary note to those students tempted to rely too heavily on unfiltered mainstream media for their scientific knowledge.

But discussion of the 2% begs the question of “the 98%” (i.e. the amount of the world’s emissions remaining once ICT’s 2% contribution is removed). What impact can ICT have on this? Is it even possible that we could “spend to save” in carbon terms? It is worth exploring the implications of this further. A good starting point would be the authoritative Smart 2020 report (The Climate Group 2008), which identifies 7.8Gt of plausible ICT-enabled carbon savings by 2020, a full 15% of current global emissions.

5. How to embed sustainable IT into a computer science syllabus

Given the importance of sustainability, both in general terms and in terms of its impact on the ICT industry, I would argue that it is crucial to embed the topic into our IT teaching. It is important to consider the character and motivations of the computer science student: typically thought to be numerate, analytical and technology-driven. Yet nowadays students approach IT from a wide range of perspectives, from electrical engineering to psychology. What follows, then, is a selection of ideas that illustrate the wide range of options available to tailor sustainability teaching to the module and to the student.

Firstly, we consider the 2% (and rising) contribution of the ICT industry to worldwide carbon emissions. There are many ways the ICT industry can and should improve its impact, through energy efficiency, least materials design and innovative new products. One important notion is Design for the Environment (United States Environmental Protection Agency, 2002) which is a long standing engineering principle, encompassing such important techniques as comparative risk analysis, cost and performance tradeoffs, evaluation of process innovations and pollution prevention measures. All of these are of clear relevance to the IT industry and will appeal to students with an engineering bent.

Looking more broadly, there is a whole field of environmental analysis called lifecycle assessment (LCA) (ISO, 2006). At a top level, this is a very simple concept; we look at a product lifecycle, from mineral (etc) extraction through to production, use and disposal (or recycling). Detailed work is needed to measure the environmental impacts at each stage (for example carbon emissions, water use, or eutrophication). There are national and international databases that help us measure this, covering things like carbon emissions from electricity (this varies by region and generation technology). But LCA is not cut and dried by any means; indeed the methodologies are a topic of intense research, with novel techniques being proposed to deal with uncertainty, assumptions and temporal differences between datasets. As an examplar, there is an issue with paper, which is often the most significant component of a typical office laser printer’s lifetime carbon impact. Yet paper could actually be seen to be sequestering, or storing, carbon that would otherwise be released to the atmosphere. LCA practitioners differ on whether to allow this sequestration in paper’s carbon impact.

An interesting technical approach relies on exergy, a thermodynamic concept that is roughly the opposite of entropy, and measures (informally) the quality of energy. The second law of thermodynamics states that although energy cannot be destroyed, it is converted from a higher quality to a lower quality form and so exergy always decreases. This gives a design metric for sustainability; can we create a system where exergy reduction is kept to a minimum? (Lettieri, Hannemann, Carey, & Shah, 2009).

This type of thinking is what we might call “eco efficiency”; that is, using less material, producing less waste, and using energy more frugally. A good example of this is dynamic smart cooling (Bash, Patel, & Sharma, 2006), which seeks to reduce the typically exorbitant costs of data centre cooling (often as high as the energy needed to actually power the servers). Such technologies thus seek to address what I have called “the 2%”.

Another approach is to look at the “98%” opportunity. One such topic of interest to computer scientists might be enterprise information management. “If you can’t measure it, you can’t manage it” is a quote much in vogue. The so-called “carbon dashboard” market is currently exploding, with offerings from startups like Greenstone and Hara jostling for market share with established providers like SAS and SAP (Verdantix, 2009). Building such a system requires a sound understanding of key computer science concepts in architecture, security, information management, databases and visualization.

This last theme brings us into the realm of the HCI practitioner. The communication of sustainability information is a key issue, whether we consider carbon dashboards, home energy metering, or community displays. There are some interesting ways to use graphics, for example the carbon quilt⁴ and Arlene Birt’s work on food packaging and lifecycle communication (Birt 2006). There are new technologies such as real time displays in community buildings and systems like i-measure from the Environmental Change Institute make use of social dynamics, such as competition between neighbours on a street to minimize their energy use (Bottrill, 2007). Thus, sustainability offers a wide range of teaching possibilities from hardware design to social engineering; encompassing just about all computer science perspectives.

These ideas are currently being used to inform the design of sustainability units on the Innovation and Technology Management MSc at Bath University. Student engagement and feedback will be captured and used to iteratively develop these themes.

3. How to engage debate

So how should sustainability debate be encouraged within a computer science syllabus? As we have seen, sustainability is certainly a topic with many promising areas for debate. However, not all debate is illuminating, and some may generate more heat than light. Climate science, for example, is a particularly divisive issue, with the popular media taking increasingly polarized positions and scandals like the data mishandling at UEA (Pearce, 2010) leaving many students somewhat confused.

There is, of course, a place to pick apart the issues, but I would advocate that computer science is not that place. Instead, interested students should be directed to reliable sources such as Real Climate⁵ and journals like Nature and New Scientist. In addition, one would expect that our sister departments with environmental science courses will deal with such topics with considerable depth and rigour. But on a computer science course such arguments are more likely to derail more relevant discussions. Certainly, in my experience working with the corporate sustainability team in Hewlett-Packard, we found it necessary to use techniques that circumvented time-consuming and unproductive debate. One approach we took was to concentrate on the energy gap and legislative drivers. Another possibility is drawn from my teaching experience in a different field, evolutionary computing, where there is a parallel issue with creationist minded students. The solution here is simple; treat evolutionary computing (merely) as a useful algorithm, thus neatly sidestepping any potential conflict. Similarly with climate science, one approach with so-called climate skeptics/climate deniers is to point out that the ICT industry (and every other industry) is adapting to a low carbon economy, and that it behoves us to be prepared for this change, regardless of whether we accept the underlying science.

Consideration of a low carbon economy of course begs the question of whether such a future is inevitable and what it may look like. Opinions vary; on the one hand we have the pessimistic outlook of Georgescu-Roegen (Georgescu-Roegen, 1975) with its vision of an inevitable reversion to a “berry picking” culture. At the other end of the scale, we have the extraordinary assertion by Gilder that a knowledge economy is anti-entropic and hence is not subject to these physical limitations (Gilder, 1981). A good place to start is, perhaps, the recent “Climate Futures” report (Forum for the Future, 2009), which presents various scenarios capturing the predictions, warnings, hopes and fears expressed by a wide range of industry experts. ICT has a particularly important role to pay in the “Service Economy” scenario, some aspects of which are already occurring (Mines, 2008).

Students should be encouraged to think about innovative ways of building a service or of meeting a need in the low carbon economy. Example might include travel replacement technologies like high-end video conferencing, paper reduction technologies like personalised or on demand digital printing, and data centre consolidation through virtualisation. Of course, it is entirely possible that this discussion will yield entirely new possibilities; their viability and potential can and should be vigorously debated.

It is almost impossible to discuss sustainability without at least touching on ethics. Although the Brundtland definition seems clear enough at first glance, it actually brings in key issues of moral judgement, which have been debated by philosophers since (at least) the time of Bentham (Bentham, 1789), who is widely associated with a philosophical school of thought called utilitarianism. The key issue to be debated here is the correct definition of utility (a computer scientist might usefully equate this roughly to an objective function). For example, is utility wealth, health, happiness, or some more complex measure? Are we attempting to optimise average or minimum utility? The theme can be further explored using the work of Rawls (Rawls, 1971), who has explored the idea of intergenerational justice. Is it justifiable, for example, to discount the ‘utility’ of future generations? This is very relevant to understanding the Stern report (Stern 2006). Stern himself, while allowing limited discounting, commented that ’we found no persuasive arguments to discriminate on the basis of birth dates’ (Stern 2007). Mendelsohn however counters that ‘a low discount rate makes every generation indebted to future generations’ (Mendelsohn 2007). Fred Hirsch (Hirsch, 1977) has provided another link to economics with his notion of moral re-entry, in which economic agents are compelled, through a combination of religious, cultural, ethical and social norms, to co-operate in order to reduce the positional competition for resources (economists would call this a negative externality). An appreciation of such topics is becoming essential to the modern computer scientist.

Currently, the idea of carbon markets is a hot topic in economics, not least because the EU Emissions Trading Scheme and the UK Carbon Reduction Commitment have made carbon trading a reality. The debate between regulation, taxes, carbon trading, offsets and other mechanisms is a debate of which computer scientists should be aware, and courses with an emphasis on IT strategy and policy should certainly deal with these topics in depth.

A promising area of debate is that of methodologies, and ways to deal with context and uncertainty. There is much literature on the subject, from life cycle analysis (ISO, 2006) to information architecture (Cayzer & Preist, 2010) and behavioural studies (Sterman & Booth Sweeney, 2007). The issue is essentially that sustainability metrics, like carbon footprints, cannot be properly interpreted without knowing a great deal about their context; the scope, estimates made, assumptions taken for granted, reference data used and so on. Such issues need to be resolved before proposals like eco labelling, currently applied in an ad-hoc way to ICT products like Hewlett-Packard printers⁶ can be standardized.

7. Links to other departments

In this era of inter-departmental alignment, it is always worth identifying opportunities for partnerships and cross-links. Sustainability offers a myriad of topics which are interdisciplinary in nature. Philosophy departments will doubtless be keen to tackle the ethics of sustainability. Economists have tools to analyze the relative merits of carbon taxes and trading, together with measures of environmental progress (subjective well being). Sustainability governance is of obvious relevance to politics courses, while climate science is dealt with by physicists, geographers and others. Finally social dynamics and behaviour change are essential tools for tackling sustainability challenges, and hence offer a chance to build a bridge with the social sciences (especially psychologists).

It is essential that computer science courses deal with these topics appropriately. To take one simple example, it is important to know that “carbon emissions” is a proxy term for a number of chemicals (not just carbon dioxide) that have an adverse effect on climate (so called Greenhouse Gas emissions). Similarly, some knowledge of economic metrics like GDP - and their limitations when thinking about sustainability — seems relevant to any assessment of low carbon technologies. These issues should prove to be an excellent bridging topic for related courses in other departments. They provide opportunities for departments to collaborate at a number of levels: visiting lectures; reuse of modules across departments; jointly developed units. In fact, in some cases there may be a case for entirely new interdisciplinary awards, such as a Masters degree in sustainability. Such a proposition is strengthened by (and in turn will nurture) a sustainability research theme that leverages multiple centres of excellence in a university. Thus, in addition to providing a series of local improvements, sustainability can provide a unifying platform that will benefit an entire higher education establishment. This is already happening in some places; for example the University of the West of England has an Institute for Sustainability, Health and Environment and the University of Bath has an Institute for Sustainable Energy and the Environment, both cross-university centres.

8. Conclusion

In this position paper I have talked about Sustainable IT and why it is important to incorporate it in our teaching. The notion of the 2% and the 98% is a useful way to think about the challenges and opportunities of sustainability for the ICT industry. Sustainability can be incorporated into the IT syllabus at a number of points, ranging from electronic engineering right through to social dynamics and behavioural change. Not all debate is useful and illuminating; in particular climate science and the likely future effects of global warming are not, I would argue, suitable topics to explore in depth in the IT curriculum. Instead, students should be encouraged to think about innovative approaches to product design, IT services that address the 98%, and robust methodologies for sustainability metrics. Interdisciplinary topics present an exciting opportunity to leverage teaching and research strengths across multiple disciplines. Sustainability is indeed a global challenge and one in which the ICT industry has the potential to play a key role. We should equip our students with the skills they need to make this happen.

Notes

¹ http://www.energystar.gov/

² http://ec.europa.eu/environment/climat/emission/index_en.htm

³ http://www.decc.gov.uk/en/content/cms/what_we_do/lc_uk/crc/crc.aspx

⁴ http://carbonquilt.org

⁵ http://www.realclimate.org/

⁶ http://www.hp.com/hpinfo/globalcitizenship/environment/products/ecohighlights-label.html