Big data’s Big problem
The problem big data doesn’t talk about, and how we’re taking small steps to tackle our own #ForABetterInternet #WorldEarthDay
Reading this, it’s likely you’ve sent or received several already today, maybe ten, a hundred maybe more? No, not COVID-19 germs. I’m talking about emails. Perhaps you have been working from ‘the cloud’, uploading files or sending large files? The last thing you’re thinking about is the carbon footprint of that activity? Maybe you should…
Emails are a big problem. Did you know an email with large attachment generates 50 grams of CO2 from the power required to send and receive it? That’s not much, right? Wrong. It adds up when you consider we send 281 billion emails a day.
Don’t take my word for it. This is information was available back in 2011, from Mike Berners-Lee: “How Bad are Bananas?: The Carbon Footprint of Everything”. Mike Burners-Lee is the brother of Tim Burners-Lee, (who invented the world wide web). I trust Mike got his sums correct .
The average full-time officer worker will send/receive 140 emails a day. Which is 34,300 emails a year creating 168kg of CO2. This is the same as if you had used 16,800 plastic carrier bags, or drunk 2,100 disposable cups of tea or coffee with a serving of milk or or drove 237 miles in an average car. (Source: cwjobs.co.uk)
Back in 2011, Mike Berners-Lee estimated that (on average) a year of emails adds up to around 135kg of CO2 emissions every year. The same as driving a combustion engine car 200 miles. That was 2011. That’s just email.
The launch of 5G, the endless introduction of IoT devices, cryptocurrency and now ‘the cloud’ is set to only compound the problem. As more of us and our devices become connected, more data needs to be processed. Processing requires energy to power the servers and water to cool the systems that drive the processors.
Electricity bill? Shocking
Think about your electric bill at home — then imagine what a bill for 10,000 servers, operating 24/7 is likely to be!
Data centres are havens for ‘dormant pollution’, ie. stored e-mails make servers run uninterrupted despite not being actively used.
In 2016, it was reported that the world’s data centres used more than Great Britain’s total electricity consumption — 416.2 terawatt hours, significantly higher than the UK’s 300 terawatt hours. Equalling three percent of the global electricity supply and accounting for about two percent of total greenhouse gas emissions, data centres have the same carbon footprint as the aviation industry.
Solar and wind to the rescue?
But wait, you’ve seen these centres in the desert in the US, and they’ve got solar-power, right? Correct, using renewable energy is under way but has much, much further to go to offset the growth in internet traffic, experts say.
Even if the industry were able to shift to 100 per cent renewable electricity, the volume of energy they would need would put intolerable pressure on the world’s power systems. Source: Independent.co.uk
By 2025 data centres could consume a fifth of global electricity. By 2040, storing digital data is set to create 14% of the world’s emissions, that’s around the same proportion as the entirety of US does today. Considering only half of the world’s population is connected to the internet, the numbers are only set to go one way, spoiler: UP.
Did you know?
In the US, an average piece of digital data (mail, download, video, web request. Source: ADEME) travels 15,000 km before reaching your screen. Why? The price of electricity. Companies capitalise by locating their data centre facilities in places where the cheapest electricity is available. Since data travels so fast, there’s no real incentive to place data centres close to where people might need them. That’s a lot of screen-miles. Do we get screen-mile points?
A 2016 Greenpeace report reaffirmed this:
“The transition to the cloud could in fact increase the demand for coal and other fossil fuels despite significant gains in energy efficiency and adoption of a commitment to 100% renewable energy because of the dramatic growth in new data centre construction by cloud and colocation companies such as AWS and Digital Realty in Virginia and other hot spots that have some of the lowest percentages of renewable electricity in the U.S.”
Thirsty? Try 360,000 gallons of water a day
Computers, servers, electrical wiring and water? Well, yes. There are over 3 million data centres in the US alone and they require a lot of water.
Servers require processors which get very hot if not cooled. Water is required to cool the servers, which heat up from being in use 24/7, and indirectly from water being used in the same way at the power plants they get their electricity from.
This article from 2009 by a data centre designer and researcher at Amazon is worrying. He estimated a 15-megawatt data centre can use up to 360,000 gallons of water a day! That’s an Olympic sized swimming pool amount of water every 2 days, per data centre. Times that by three million.
Shall we mention streaming services? Your evening on the sofa, relaxing watching the latest boxset is not carbon-free, far from it. Streaming services (Netflix fans look away now) perform significantly worse. Services like HBO (Greenpeace rated D), Hulu (Greenpeace rated F), and Netflix (Greenpeace rated D) scored badly since they demand so much energy.
A Berkeley Lab Study in 2014 cited that in 2011, Americans streamed 3.2 billion hours of video (that’s a lot of “Game of Thrones” and “Breaking Bad,” folks). All that streaming consumed 25 petajoules of energy — enough to power about 175,000 U.S. households for a year — and emitted 1.3 billion kilograms of carbon dioxide.
What’s can be done? Who is doing anything about it?
A study by the International Data Corporation (IDC) claims that by 2025, worldwide data traffic will have grown by 61 percent to 175 zettabytes, meaning roughly 75 percent of the population having at least one data interaction every 18 seconds. I don’t even know what a zettabyte is, but it sounds large.
One thing is certain, we can’t live without data centres. How else will one be trolled on Twitter without them? The climate emergency is by no means a new topic — the same can be said for the sustainability of data centres.
There’s the “The European Code of Conduct for Data Centre Energy Efficiency programme — a voluntary initiative created in response to increasing energy consumption in the sector — which dates back to 2008, but has gained renewed traction with a seismic shift towards real change:
“Over the past twelve months we have witnessed a sea change in the industry’s views on Green IT and sustainability. The topic is no longer perceived as the domain of the ‘tree hugger’ or ‘eco-warrior’ but rather a fundamental element in any organisation’s strategic business plan, and IT has played a major role in this shift in attitude.” — Phil Worms, Marketing Director, iomart Hosting (source)
Companies like Facebook, Google, and Apple score well on their Greenpeace assessment. (They all ranked A in the 2016 Greenpeace Report.) They’ve set 100% renewable energy goals, they’re transparent about the process, they have facilities in renewable energy hotspots like California, and they have money to throw at the initiatives.
For example, Facebook opened a major data centre in the far north of Sweden, in the mining town of Lulea, 70 miles from the Arctic Circle, why? The 84-acre site houses tens of thousands of computer servers that still require 500 huge fans to cool them. Housing data centres in very cold climates dramatically reduces the energy needed to cool the facilities , with an abundance of readily available renewable energy.
Other data centres are taking a different approach to reduce the amount of energy they waste. Nordic data centre operator DigiPlex has made a pledge that will see the waste heat from its facility in Ulven, Oslo, reused to warm 5,000 apartments in the city. The company signed an agreement with a local district heating supplier to redistribute the heat generated by its data centre, which is also renewably powered.
Long and short of it? There’s movement, but the pace is slower than the melting glaciers their contributing towards melting, and there’s a long, long way to go.
If it is broken, fix it
I’m far from qualified and I don’t have the answers regarding big data and carbon emissions, but there are solutions being developed that some Data Centres are embracing.
In researching this article I requested information from the Data Centre company where our secure server is located. While they have signed-up to the EU Code of Conduct for Data Centres; promising to meet ‘energy efficiency best practice’, minimum procurement standards and annually report on energy consumption. Which sounds like legislative language for ‘do the minimum and make it sound more important’, like a patronising pat on the head to anyone who cares or dares to ask questions.
Our hosting provisions. We relocated our secure server to their Maidenhead branch, the nearest to our location which reduces screen-miles required to serve you our work!
To recap emails, large attachments, sending large files, streaming, data centres and the cloud — it’s an enormous drain on energy and water and generates more carbon than we were ever aware of. It’s broken and it’s a problem. Do not despair. While global companies like facebook are opening major data centre in freezing climates (to reduce the necessity for coolant systems) there are small things we can all do.
At Plan-B Studio we’ve been making small, but not insignificant changes to how we work. A new website, cleansed data, new secure servers located closer to our place of work. We’ve also taken advice which you can follow too regarding emails:
- Unsubscribing from mailing lists we never take notice of
- Removed all unwanted emails
- Cleansed our mailing list
- Cleaned our database
- Reduced the size of emails by (where possible) lowering the resolution and compressing images and avoid large HTML elements.
- Remote working (we’ve been office-free since 2009 and survived)
- Linked to files or information online rather than adding an attachment.
The last item was a lightbulb going off in my head. I’m often approached for requests to send people our ‘portfolio’; a PDF showcasing our work.
Problem is I’m a cynical perfectionist. I take great pride in ensuring our work is presented without any compression or loss of quality. This has created files too large to email.
We have used WeTransfer for years. But we’re sending more and more large PDF files (c.138mb!), sometimes several times a week. As part of our ‘small but not insignificant changes’ I challenged my team — is there a better way? Can we ‘link to files or information online rather than adding an attachment’?
The answer was a big, fat, beautiful: YES.
We’ve built our own Presentation Display Format (or PDF for short, not sure that acronym will catch-on). A tool within our Wordpress CMS containing a series of pre-designed modules (here’s an example).
Our presentation tool minimises energy requirements, it doesn’t load any additional assets or images. We select the projects we want to feature, which are scrapped from our website and hosted from our server.
The only energy generated is on-demand, when/if a recipient clicks on a project.
Each presentation is assigned a random ‘slug’ (the part that comes at the very end of a URL) which we can personalise. Allowing us to quickly create a curated selection of work, include testimonials, stories, quotes, introduction slides and a welcome message to our guest.
This way we create a little bit of personalisation online, we curate the work we want to show, in the order we want to show it. Including anecdotes, information, all contained in one simple email with a link.
I won’t save the world, but we’re always open to improvement.
