Powering small to medium size tech companies on renewable power

I’ve been thinking about the steps you can take to green the tech sector, and the more I think about it, the more I think the biggest step for most web/SaaS companies, is likely to be how they power their infrastructure, if it isn’t related to employee transport.

This is confirmed somewhat by the CSR reports by various companies mentioned in this fantastic new resource, climateActTech:

A quick way to estimate what your company’s carbon footprint might be is to find a peer who has a similar business model. At most “cloud based” technology companies the majority of the carbon footprint comes from data center energy use. At other companies it might be from fabrication and manufacturing processes, company travel or directly from other company business interests. Here are a few examples.

“Cloud Heavy” Companies

Manufacturing Companies

Consulting Companies

Knowing where the bulk of your carbon pollution comes from will help you achieve the greatest gains by directing your efforts appropriately.

The thing is, if you’re a large company like Google and Apple it’s possible to do some power purchase agreement (PPA) to effectively power your infrastructure with renewabbles, but as companies get smaller, the options (short of using Google’s own infra, for example) become more restricted too.

So, this piece in Green Biz caught my interest, as it’s the first time I’ve read about smaller companies clubbing together to do a ‘virtual PPA’ :

With FERC’s blessing, Apple and others can participate in so-called “physical” PPAs, where the contract holder actually takes ownership of the energy generated by a project and its sustainability certificates. For those without FERC clearance, a “virtual” PPA offers a chance to buy clean energy from a project at a long-term fixed price without technically being the owner of the power in play.

“Most deals, aggregated or not, are going to be virtual,” Kelly said.

How those agreements are divvied up in an aggregated deal could vary.

One model is to have each company sign its own power purchase agreement — the predominant model to date in the deals that Kelly has seen. An alternative could be an “anchor tenant model,” Kelly said, where one company signs a PPA and others agree to individual contract terms.

In the latter scenario, however, the anchor tenant would become as a middle man that must be comfortable with other participants’ credit ratings.

“It’s kind of putting its own balance sheet out there,” Kelly said.

When you have organisations like Lumenaza offering ‘virtual power stations’ by aggregrating supply from loads of micro scale power generators, and I’m now wondering how small these PPAs can realistically get.

Also, given how Feed In Tariffs in Europe are dwindling over the next two years, it feels like there will be a lot of micro scale generators looking around for new purchasers of power, to keep their plants financially sustainable.

If there are companies who are looking for green power, and providers who will increasingly be looking for new purchasers of power, surely this service must exist already, right?

Right?

 

 

Computing, Climate change and all your relationships

Generally speaking, it’s a small, but growing group of folk who seem to interested in the intersection between climate change and tech.

Earlier in May, I came across Nabil Hussein‘s talk, Computing, Climate Change and All Your Relationships, and boy is it a breath of fresh air – climate change for me is a very much an issue about fairness, and treating humans with basic decency, regardless of where in the world they are.

This is and remains on of the reasons I’m attracted to the web – as Tim Berners Lee said, it’s supposed to be a tool for everyone, not just a small group, and that human potential it allows us to unlock is a source of continuous wonder for me.

Tim Berners Lee, in the London Olympics, with his cameo for the world wide web

I’ve seen a fair few talks, but this is the first time I’ve seen someone in tech talk about climate change in terms of the sheer numbers of people who don’t go to tech conferences who are affected, and how it disproportionately affects people least able to cope with the wrenching changes it’s bringing about.

There was one slide in particular that was particularly effective here:

Screen Shot 2018-08-16 at 18.47.33.png
Comparing the Global North (14% popn, 73% income) and Global South (86% popn, 27% income)

Challenging Bret Viktor’s “options” approach

This is also the first talk I’ve seen that critically engages with Bret Victor’s essay, What Can an Technologist do about Climate Change?

We’re seeing how policy makers in the White House have been undoing progress they’ve made, and you see the same patterns in the UK, with fracking over building a local renewables.

Screen Shot 2018-08-16 at 19.00.41.png

That said, I can’t say I’m totally behind this approach. and I think, having a binary here is oversimplifying things. There’s absolutely cases where in the global North you’re seeing terrible decisions, but there are also cases where you are seeing progress being made, and that helps set the tone for future, more progressive, climate friendly policy.

For example, I see the investment by Germany in renewables over the last 20 years pretty important for decentralising how power is generated over here: both to bring down costs sufficiently to ‘create options for people’, and also counter the concentration of power among the incumbent providers of energy, and encouraing many more small scale producers of energy from renewables.

I guess the counterpoint here is that China, which (on this map at least, is included in the global south) has also played a huge part in shifting the economics of solar, but that’s hardly been a bottom-up approach. It’s been a very deliberate policy decision, the same way China’s National Sword policy was a top down decision to stop importing so much waste from the global north.

Anyway, it’s totally worth 40 minutes of your time if you have a passing interest in climate change, and you work in tech. There’s also a super handy transcript online too. Enjoy!

How much CO2 can you save when you remove ad-tracking from news sites?

Now that GPDR has landed, we’re seeing companies serving EU specific versions of their site to EU users, which in some cases, serve a user experience which is cleaner, simpler and faster loading.

Some are much smaller over the wire too. So, because moving data uses servers,and those servers use electricity, and that electricity usually comes from burning coal, I’ve had a go at doing some basic calculations to work out what the CO2 reductions might be if these became the norm.

TLDR – For a site like USAtoday, it looks like running the ‘GPDR-lite’ version as the default would represent CO2 emissions savings equivalent to an entire European persons’s annual carbon footprint, each month.

How I arrive at these numbers

I need to stress right at the beginning – these figures I’m about to share are very rough, and I don’t pretend that they’re in any final, accurate form at all.

I’m hoping to share this to help get a better idea for how you might work out the CO2 emissions associated with transferring data in a more rigorous fashion,  but also because I think these emissions are worth discussing, and I can’t find any numbers like this online yet.

First, our smaller site:

Let’s take Hadley’s tweet here – she’s referring to a the EU specific version of the USA Today site, which is about 500kb, compared to the full size, ads and tracking site which weighs in at 5mb site instead:

What kind of energy footprint does this represent? Let’s take a rough estimate of the total daily traffic for the USAToday from EasyCounter (we could use Alexa if we wanted to pay for more accurate traffic results)

Screen Shot 2018-05-27 at 16.45.50.png

Looking here, we get 3.77 million daily page views.

So, if we wanted an idea of daily bandwidth, we might multiply page size by daily page views.

So, lets take a 5MB page, and multiply it 3.77 million times to represent the page views.

This gives us a figure, that, when we round it to the nearest one hundred gigabytes, is about 18,400 gigabytes of data per day.

How much energy is that?

Now, it’s been really hard to find reliable numbers to convert bandwidth to energy usage when I’ve looked before, but the best freely available figures I’ve found are from work I found via Jonathan Koomey’s blog, where he shares something like this:

This article derives criteria to identify accurate estimates over time and provides a new estimate of 0.06 kWh/GB for 2015. By retroactively applying our criteria to existing studies, we were able to determine that the electricity intensity of data transmission (core and fixed-line access networks) has decreased by half approximately every 2 years since 2000 (for developed countries),

So, this is semi-throwaway blog post and it’s a Sunday, and so for the purposes of getting a ballpark figure, I’m going to cheat and just project forward two years to 2017, and say 2018 is close enough to 2017 for me to use 0.03 kWh/GB.

Okay, how much carbon dioxide is that?

For a ballpark figure like this, we’d take the total energy needed to transfer our 18,400 gigabytes per day, then multiply that by our 0.03 kilowatt hours per gigabyte, then multiply that by the CO2 emissions per kilowatt hour.

Let’s get our CO2 per kilowatt hour figure so we can do this

In the US, where most of the USAToday audience is likely to be, a fair amount of coal is used to generate power, so when I was dumping some numbers into this jupyter notebook, I made a guesstimate figure of 0.45 kilograms of CO2 emitted per kilowatt hour of electricity generated.

After checking it against the emissions index website to check against some actual numbers, it turns out I was off, but not that far off.

Their figure is 432 kilograms per megawatt hour, which is about 0.43 kilograms of CO2 per kilowatt hour.

Screen Shot 2018-05-27 at 16.56.42.png

I’ve been rounding at each stage here, to make it a bit easier to keep the numbers a bit easier to remember.

However, I’m going to pull the actual number from the jupyter notebook, which when rounded again, gives us 248.5kg of CO2 per day, or a little under 7.5 tonnes per month more or less.

Translating this into something  more tangible

If we had the lighter, GPDR friendly, ad-and-tracking-free version of the site as the norm, if we just looked at the bandwidth savings, then we’d be saving something like the annual carbon footprint of a typical European, according to the World Bank, every month.

Or if you prefer, something like a flight between New York and Chicago every day.

If this is interesting to you, there’s some more in this jupyter notebook on github.

Plug time – the planet friendly web guide

I’m looking for people to work with and explore this kind of stuff with me.

Why?

Well, I work in tech, and it seems like loads of our existing tools, and practices can be re-purposed to bring about reductions in CO2 emissions in our industry AND make the digital products we create work better for our users.

If you design infra for services,where you source power, and how you provision your resources, to match your use (i.e. scaling) has an impact.

If you design clients, or apps, then how send data over the wire has an impact (i.e. WPO, et al).

If you design the business model, or how you get feedback from stakeholders, or how you travel to do all this, then decisions you make here have an impact too.

Let’s chat

If this is interesting, please do get in touch, as I’m trying to:

  1. get a free, open source guide together around the subject at planetfriendlyweb.org
  2. prototype a workshop (one remote, one in Berlin), to help organisations identify where they can make these reductions (and usually, save money or reduce waste along the way)

My contact page lists the usual ways to reach me, and if the guide seems fun, there’s a contributor page there too. T

 

 

 

 

 

My talk about a planet friendly web at DataNatives in Berlin last night

So, last night, I did a talk at Big Data Berlin, with the shockingly linkbaity title of Green Clouds in a world of Blockchains and AI.

Rather impressively the video is online already – and you can see it start at 37:01 in the video linked below:

https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2FDataNativesCon%2Fvideos%2F2032555460290623%2F&show_text=0&width=560

The talk

The talk is online on speakerdeck, like my other talks:

The reception

Generally I was really pleased with the reception here – the audience (a mix of data focussed lot, but also a devs, designers, and biz poeople) was really engaged through the 20 minutes I took to deliver it, and I got laughs at all the places I was hoping.

However, you won’t hear any of them on the video above, as the sound recording seems to be coming from the mike, and it sounds like tumbleweed whenever there’s a pause for people laughing.

Anyway, I hope the content is interesting, and the links are all in the deck linked above.

If the stuff I was talking about interests you

When I find the time, I’m trying to get a guide together, for people who build digital products and services, who want to make them greener. You can see it at https://planetfriendlyweb.org, and there a loads of ways you can contribute too – check the contributors page to see.

Also, feel free to drop me a message via the contact page, or leave a comment below.

Ta!

How much of the web runs on renewables today?

As part of the work I’m doing on the Planet Friendly Web, I’m trying to get access to data that I can base the guide on. In some cases this involves creating datasets from existing data. Here I share some findings from a dataset I generated along the way.

For example, to get a figure on how much of the web runs on renewable power, I started with a dataset of the top 1 million domains by traffic from Alexa.com, then run the list against the Green Web Foundation’s own API, which maintains a list of which domains run on renewable power.

To do this, involves making something like 100k API requests, so I created a screenscraper to carry out the job, and take care of retries, failed requests and so on. You can see it here on github.

I’ve uploaded the dataset created to datbase, partly as an experiment in making it available in a decentralised way, but also partly try out the workflow for publishing data.

So, now we have some data, let’s see what we can do with it, right?

Doing some analysis and some interesting findings

I have an earlier exploration of the data in a notebook on github, but when working with this data, I ‘m bit embarrassed to say I forgot how to use the Dataframe filters to slice the data quickly.

So instead, I’ve used Open Refine. You could probably store this in a Google spreadsheet too, as 100k rows is big, not but THAT big.

Anyway, what do we see?

There’s a few interesting findings just from faceting data like below in Openrefine,  and sorting by count along a few dimensions:

Screen Shot 2018-05-15 at 12.07.36.png

If you’re not familiar with OpenRefine, I’ll summarise what’s visible in this view:

  • Youtube.com is now more popular than google.com. Who knew?
  • The top three websites in the world run on renewable power. Huzzah!
  • Based on the greenweb foundation’s data, around 7% of the web the most popular domains on the net run on renewable power.
  • Hetzner AG, a German hosting company hosts more domains running on green power than Google does.
  • Amazon doesn’t appear here at all as a green provider.

After a slow start, I understood Amazon to be a HUGE player here, and while they have a nice shiny page showing off their windfarms and how much renewable power they use , they also run a load of their servers on coal. That they don’t appear may be an artefact of the Green Web Foundation going by an organisation’s entire power mix, to decide whether a company is running on green power or not.

I think need to check with Rene at the Green Web Foundation to see.

Fancy playing too? Come hang out on slack

This shows some pretty superficial analysis, but there’s already some interesting nuggets here.

If working with this data sounds interesting to you, let me know in the comments – I’m looking for collaborators on the Planet Friendly Web Guide.

Alternatively, come hang out in the sustainableux.com slack channel, where there’s a nice little community growing around sustainable web design.

If you prefer email

It turns out there’s a W3C Sustainable web design group. Here’s my post to the mailing list, if you’d prefer to communicate there via email.

 

 

I’m hosting the Mozilla Global Sprints in Berlin in May – come along!

I’m helping host the Berlin Mozilla Global Sprint next week – it’s a two day event, set aside to create the space to make it easy to volunteer on existing open source projects, aligned with the key Mozilla’s key Internet Health issues, outlined in their recently published Internet Health Report.

More specifically, these issues, taken from the report are:

WEB LITERACY: Projects that teach individuals skills to shape — and not simply consume — the web.

OPENNESS: Projects that keep the web transparent and understandable, allow anyone to invent online without asking permission, and encourage thoughtful sharing and reuse of data, code, and ideas.

PRIVACY & SECURITY: Projects that illuminate what happens to our personal data online, and how to make the Internet safer for all.

DIGITAL INCLUSION: Projects that ensure everyone has an equal opportunity to access the Internet, and can use it to improve their lives and societies.

DECENTRALIZATION: Projects that protect and secure an Internet controlled by many, so that no one actor can own it or control it or switch it off.

Oh neat, these are things I’m totally in favour of. What projects are there?

There’s a load of open source projects you can hack on listed at Mozilla’s Pulse page. But to be honest, as long as you’re working on a project that addresses the issues listed above, you’ll be welcome.

If you’re feeling particularly generous, and heroic

I’m looking for some help on a project called the Planet Friendly Web Guide, which I was working on last earlier last year, as part of the Mozilla Open Web Leader programme.

I presented it at SustainableUX, and if you’re visually inclined, you can see the deck below I that I used:

I’m looking for help in a bunch of ways, but the simplest way to see where you can help is to visit this contributing page on the guide.

In particular, I’m looking for help building some fun little widgets to let people get a quick idea of the carbon footprint of the infrastructure used to serve sites they use or build, based on the general platform, packets, process model, to see if it’s easy to apply for someone who hasn’t been working on the project like I have.

Come, and hack on something nice

So, to recap, the deal is basically:

  • turn up
  • hack on a thing that largely agrees with the principles outlined in Mozilla’s 5 key issues
  • be fed at lunchtime as a token of appreciation if you’re giving your time to make the web a better place

There’s a lot more about the whole idea of Mozilla’s global sprints on their dedicated site.

You can sign up here on the registration site.

Wait two whole days? In this weather?

It’s also totally cool to drop by for just part of the two day period – understandable if you just want to spend a bit of time on a project, before getting out and enjoying the wonderful Berlin summer weather.

You in?

 

How much CO2 does an office worker generate per year?

I just posted this to friends on Facebook, and it seems a good idea to share it here too, to help with my search:
Hello internet friends. Would a kind soul be able to help me out here?
I’m doing a recorded 20 minute talk in Feb about the environmental impact of building digital products for a free online conference, http://www.sustainableux.com, and I’m looking for leads to work out a number I want to refer to in the talk.
 
I’m after a single number, for the average carbon footprint of a single office worker, working in an office, full time, including their commute, in terms of tonnes of CO2 per year.
 
I know commutes vary wildly, and that’s fine. For this, an average will suffice, as I’m not pretending this will be accurate, just a ballpark figure.
 
I think I’ll mainly be speaking to an audience based in the US, or Europe, based on last year’s viewer figures, so extra points if the number applies to those regions.
 
I know this figure won’t be precise, and I’m aware there are loads of factors that affect this anyway. To the nearest tonne is probably okay.
 
If it helps, please think of this as a number for all those other places that aren’t where you work. I appreciate your office might be super green and virtuous and you’d love to tell everyone about how much you recycle, and how you’re going to eco-heaven, and how everyone is else is terrible but for the purposes of this request I don’t think it will add to the conversation here.
 
Sorry about sounding grumpy in that last para. I’ll be super grateful and give you a shout out in the talk if you can find it.
 
Thanks y’all ❤