fredag 1 mars 2019

The cloud devours electricity (radio interview)

.
I was on Swedish radio yesterday, in the weekly science/sustainability show "Klotet"(The Globe). The theme of this episode was The cloud devours electricity - can we store data sustainably? ("Molnet slukar el - går det att lagra data hållbart?"). The show was built up around two pre-recoded reportages and they also had two invited guests who commented and discussed the intersection of on the one hand computing and social media use and on the other hand sustainability. Besides me, the other guest was professor of Communication Systems Erik Agrell from Chalmers University of Technology in Gothenburg. I had to go to Uppsala to record the show so it took half a day (plus preparations!). The chat/interview with me and Erik took an hour but was later cut down to 20 minutes or so.

Sustainability in this context primarily means energy use and related carbon emissions, but it could also include material throughput (us buying and throwing away gadgets or data centers buying and throwing away servers). I took the partly unspecific and unstructured questions I got sent over by mail beforehand and structured them a bit so here's what they wanted us to comment/talk about:

- On the electricity consumption of data traffic in relation to sustainability:
   - Is this a big problem now and/or in the future? If so, how do we solve it?
   - Do people need to change their behaviour (just as we need to change how we transport ourselves etc.)?
   - How far can (future) technical developments take us?

It just so happens that I am giving a ph.d. course this term (together with my colleague Miriam Börjesson Rivera) and several of these questions relate to issues we discussed at the previous seminar or that we will discuss at the next seminar. So I prepared by leafing through paper I had just read and by reading papers we will discuss at our upcoming seminar so that was pretty convenient.

I knew that the focus of both the radio program and that of Erik Agrell would be on technical challenges in terms of how much the data traffic grows each year (≈ 25%), how much global electricity generation grows each year (≈ 3%), how much technical developments on energy efficiency might counteract such trends etc., so I decided to try to open up the discussion to take into account also other factors instead of focusing only on such technical factors (which are not my forte). This part of my argument primarily builds on the arguments made in this 2016 paper:

Preist, C., Schien, D., & Blevis, E. (2016). Understanding and mitigating the effects of device and cloud service design decisions on the environmental footprint of digital infrastructure. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 1324-1337). ACM.

Priest et al. talks about "the cornucopian paradigm" where:
1) "Effectively, the provision of digital services to high-end users stimulates latent demand in mainstream users for such services".
2) "The mainstream users then pick up devices and services which were formerly high-end, and they become embedded in everyday practice. Services that most users were happy without become essential to everyday life for the majority of the populace in developed countries."
3) "This results in a reinforcing feedback loop encouraging growth of the digital infrastructure."

A growing digital infrastructure then enables the design of new high-end services that later/again makes increased demands on infrastructure (and so on). This is succinctly summarised by an image in their paper:


The paper also has a long list of design principles that drives the increase in infrastructure demand that I found useful when I exemplified and pretended that I used a GoPro camera to record the daily commute to my job. Such a video stream could be automatically backed up to the cloud and I might also make "unreasonable" demands at a later point in time, for example assuming that I will have instant access to this video stream and can choose to share it a year later when I'm on vacation in Antarctica (filmed in high-definition video of course).

Basically I wanted to emphasise that the infrastructure does not grow autonomously but rather because new technical possibilities (increased storage, increased bandwidth, new capabilities and features in the electronic gadgets we buy and in the software that runs them) creates the possibilities for developing new high-end data-intensive services and because we as users are quick to pick up those services and incorporate them into our behaviours and into our lives. This line of reasoning problematises and opens up discussions about our behaviour. Are there limits to our needs (or our "needs", e.g. wants)? We only have 24 hours per day, but we can replace less data-intensive practices with more data-intensive practices, including but not limited to using more than one device/service simultaneously.

This line of reasoning analyses the problem, but a second article starts the discussion about what to do to "solve" this problem. That article is written by Kelly Widdicks and me and it's currently in review. It's a kind of continuation of an article we wrote together a bit more than a year ago, "Undesigning the Internet: An exploratory study of reducing everyday Internet connectivity". That paper was written together with yet three more persons and it was presented at the Fifth International Conference on ICT for Sustainability (ICT4S) in Toronto a bit less than a year ago. The new, just-submitted article that we are writing, "Breaking the Cornucopian Paradigm: Towards Moderate Internet Use in Everyday Life" has just been submitted to this year's (upcoming) workshop on Computing within Limits. Borrowing from another area, here then is the line of reasoning I made in the radio show:

We should all switch to buying ecological food, but it costs more so many people don't. It could however be the case that many buy ecological food not in order to "save the environment" but because they believe it is more healthy for themselves and because they care about their children's health. So these two motivations pull in the same direction and it might be that the latter (health) has more pull than the former (sustainability).

It might similarly be an uphill battle to try to convince people to use their phones less in order to "save the environment", but there might exist strong non-sustainbility reasons for why we should think about tempering our use of electronic gadgets and these drivers could also have beneficial effects in terms of sustainability - so we should explore which they are. We have four examples (reasons) in the paper but I'll just mention one here and that is "Relationships". It might be beneficial for our close relationships if we put our phones/gadgets away more often; We might ban them from the second floor in the house (where our bedrooms are) or from the dinner table (to encourage unmediated face to face conversations). Some schools have "smartphone hotels" where you park your phone, perhaps we should have smartphone hotels also in our homes? Perhaps "parental controls" should be extended to cover the whole family's use of electronic gadgets in the home?

I also cursorily raised the questions of etiquette in the radio show. There are etiquette rules in many different areas of life but few around our use of electronic gadgets. Perhaps we should speed up the invention of such rules? For a fascinating take on the evolution of etiquette, do see my blog post about Norbert Elias' (1939) book "The Civilizing Process",  a book that in detail analyses the evolution of etiquette rules over centuries (do also look the quotes further down in that blog post).

All in all it was fun to be on radio. I also thought a bit about what I could have done different so that my next appearance on radio will be better. Lastly I also suggested two other topics for radio shows so don't be surprised if I write a new blog post about my next appearance on radio a year or so from now.
.