Talk:Sustainability Initiative/Archives

From Meta, a Wikimedia project coordination wiki

This is an index of archives; for recent discussion, see Talk:Sustainability Initiative. Please do not post new comments on these pages; if you wish to revive a discussion, either move it back to the main page or link to it. For help archiving pages, see standard archival system .

Past efforts[edit]

Past related discussions[edit]

I'd imagine there are discussions on the larger wikis (German Wikipedia, English Wikipedia, etc.) and elsewhere on Meta-Wiki as well. --MZMcBride (talk) 04:40, 31 March 2016 (UTC)[reply]

Hi MZMcBride, thank you for the links! I have already looked at past discussions of this issue on German and English Wikipedia and left notes on the talk pages of the participating users (Example). Also, I informed the relevant portals in German, English, French, and Italian Wikipedia. Also, there has already been a discussion about this essay on Swedish Wikipedia. --Gnom (talk) 07:44, 31 March 2016 (UTC)[reply]
No problem. I think a "Related reading" section similar to Endowment#Related reading can be really useful for providing historical context and for researching important issues that were raised in past discussions. We have lots of discussion venues/forums in the Wikimedia universe, so aggregating links to the various threads is helpful, in my opinion. --MZMcBride (talk) 23:48, 31 March 2016 (UTC)[reply]
Why not, MZMcBride! Which links do you think are relevant? Thanks, --Gnom (talk) 09:27, 1 April 2016 (UTC)[reply]

Discussions about whether we need to do this (climate change, Wikipedia as a source of information about the environment...)[edit]

COP21[edit]

Should probably also mention COP21 outcome to show the universal consensus on the need. There's even a http://www.iigcc.org/ Nemo 08:12, 23 February 2016 (UTC)[reply]

No alternative view? Wikipedia is environmentally positive.[edit]

There is no way provided to advance the (perhaps controversial) view that (at least some of) the measures advocated by the essay are poppycock. We can only support the essay or keep quiet.

I disagree fundamentally that the impact of Wikipedia is negative environmentally. It seems trivially obvious to me that the amount of time, effort and even fossil fuels saved by having this collection of information in one place easily available to all means that Wikipedia has a positive effect on the environment.

Psb777 (talk) 22:40, 24 March 2016 (UTC)[reply]

Hi Psb777, thank you for weighing in! Which ideas do you disagree with, and why?
My assertion that WP is environmentally positive is based on the fact(?) that it used to be much more expensive to look something up. Without the likes of WP today I would have to visit more web pages before I found a good explanation of, say Bayes' Theorem, or the area of Argentina. Less energy consumed, therefore. Without the likes of the Web I would have to telephone friend or visit a library. Each less environmentally friendly, I suggest, per unit of information transferred. Psb777 (talk) 15:32, 12 February 2017 (UTC)[reply]
Assuming that Wikipedia as such might even have a positive impact on the environment to begin with, shouldn't we try to do even better by causing even less carbon emissions? --Gnom (talk) 23:00, 24 March 2016 (UTC)[reply]
Yes, and I should take more exercise, and go to church occasionally. Of course things can be more positive than positive but the notion I criticise is that WP might be environmentally negative. Only in the sense that the Enlightenment might be environmentally negative, that human beings are environmentally negative, only in the sense that to save the planet I'm unsure if you should respond. That might be bad, environmentally. Psb777 (talk) 15:32, 12 February 2017 (UTC)[reply]
Hi Psb777, I agree with you that the very existence of Wikipedia can be viewed as a positive thing for the environment, especially if we think as a world in which Wikipedia does not exists and there are paper encyclopedias in libraries all around the world... Nevertheless, the purpose of the essay is pretty clear: it just says that Wikipedia is a huge website and that several things could be done to lower the carbon emission for this website. I'd say that this is pretty uncontroversial, don't you think?. Aubrey (talk) 18:29, 27 March 2016 (UTC)[reply]

Contributions to the environment: [1]. :) Nemo 18:32, 9 June 2016 (UTC)[reply]

Net positive?[edit]

I agree to some extent with the statement immediately above... the result of making information accessible easily online should have some offsetting value against the carbon footprint of running Wikimedia projects and servers. How much energy is saved in transportation costs, by making Wikimedia resources so readily accessible? Another contribution toward a potential net positive environmental impact outcome is the provision of information regarding environmental topics... which stand to inform and support people wishing to undertake their own environmental projects or studies which can also lead to positive environmental outcomes. I have contributed to many of these articles on Wikipedia in recent years, refer back to the frequently, and am forever sharing Wikipedia links regarding environmental subjects with my peers. --Danimations (talk) 01:24, 25 March 2016 (UTC)[reply]

Hi Danimations, I doubt we can calculate the environmental impact of Wikipedia outside of the energy used by simply running it. But even if Wikipedia has an overall positive effect on the environment, wouldn't it be great to have the servers run on renewable energy? --Gnom (talk) 19:27, 25 March 2016 (UTC)[reply]
Hi Danimations. I agree with your points, Wikipedia is an important source of reference and surely it has an impact in outreach and environmental discourse and learning. Still, figures are hard to get for this kind of impact. I'd love to see something more "scientific" about what you say, but at the present moment I think that we, as a community, can actually be a bit more aware of the real immpact that the servers are having on the environment with basic metrics as carbon footprint. I understand it's not all the picture, but is something that's true and that we can count. Aubrey (talk) 18:33, 27 March 2016 (UTC)[reply]

What is wrong?[edit]

  • Please! DO NOT FORGET that not all of renewable energy power plants cold be considered as safe for animals! (Idot (talk))
  • Huge Solar Concentrator DO KILL bird! they just blind birds so they die! (Idot (talk) 15:04, 7 February 2017 (UTC))[reply]
  • a typical Wind Power Plant DO KILL bats! they just invisible for bats sonars so they chopped and die! (Idot (talk) 15:04, 7 February 2017 (UTC))[reply]
  • There is no really benefit in powering namely Wikimedia servers from renewable power sources. Because the renewable power for now is limited worldwide, taking some of "green energy" for the Wiki is taking it away from other potential consumers. The growth of energy production from renewable sources matters, not the distribution of it.--Yellow Horror (talk) 12:54, 8 February 2017 (UTC)[reply]
Hi Yellow Horror, please note that increasing the demand in renewable energy will yield more investment in creating it rather than keeping fossil fuel power plants running. This effect is very visible in Germany, for example. --Gnom (talk) Let's make Wikipedia green! 13:53, 8 February 2017 (UTC)[reply]
Isn't that effect more related to government support like tax preferences? Isn't more effective to invest money to renewable power generation facilities directly than to just increase demand that lead to increase prices at the first place?--Yellow Horror (talk) 16:30, 8 February 2017 (UTC)[reply]
Yellow Horror, people will only invest money in renewable power generation facilities if there is a high demand for renewable energy. This is why it is important to raise this demand. --Gnom (talk) Let's make Wikipedia green! 16:52, 8 February 2017 (UTC)[reply]
Please, don't mix people with corporations. People may spend their money as they wish immediately. Corporations exist to make profit in a long term. And, as i say before, there is no specifically benefit in convincing Wiki to buy more "green power". You can go and convince your apartment power supply company to buy more "green power" with the same result. Well, almost the same - it is very possible that, if you achieve a success, you'll receive an increased bill then.--Yellow Horror (talk) 17:39, 8 February 2017 (UTC)[reply]
You may ask polar bears or the animals and the indigenous people in Yasuni National Park in Ecuador what they think about cheap energy made with oil. And you may calculate what will be more expensive in the end: some green energy or handling the climate change. NNW (talk) 17:53, 8 February 2017 (UTC)[reply]
And additionally a list of companies which exist to make profit and aim to use 100 % renewable energy: http://there100.org/companies. NNW (talk) 18:34, 8 February 2017 (UTC)[reply]
So, be an environmentalists is easy and pleasant, while someone else pays for it?--Yellow Horror (talk) 18:48, 8 February 2017 (UTC)[reply]
The one who understands that it can't go on with oil and coal and who is able to pay for it. WMF will be able to pay that tiny little bit more, believe me, for the benefit of the world climate. NNW (talk) 19:14, 8 February 2017 (UTC)[reply]
So the WMF's target is making as much profit as possible? That's sad! By now I always thought it was a non-profit-organisation. Thank you for enlightening me! KPFC💬 10:15, 10 February 2017 (UTC)[reply]
  • Remember also that some power sources counting as "renewable" really aren't. For example, tidal power-plants use the Earth rotation energy that is limited (at least, comparing to solar-related energy sources).--Yellow Horror (talk) 12:54, 8 February 2017 (UTC)[reply]
You are right. When the sun will swallow Earth its rotation will end and the sun will still shine. And the oceans will be evaporated even in 1.1 billion years from now so tidal energy is no reliable source at all. I don't think we are finished with Wikipedia in 1.1 billion years. 91.60.28.172 13:22, 8 February 2017 (UTC)[reply]
  • Ok, just two simple questions: (1) What do you think is more honest: go here and vote to make Wikimedia a 100% renewable power consumer; or go to your power supply company, ask them to buy renewable power for your household and pay the difference in price? (2) If the Wikimedia have a little extra money, what do you think is better to environment: spend them to purchase more renewable power than before; or spend them to renewable power generation directly?--Yellow Horror (talk) 19:19, 8 February 2017 (UTC)[reply]
Really, you should suggest WMF to create renewable energy. Wikidata is more than four years old, it's time for a new project. Or suggest to shut down all servers. We could save so much energy. And time. And nerves. NNW (talk) 20:25, 8 February 2017 (UTC)[reply]
If the WMF buys renewable energy the companies who are offering it will spend the money in generating new renewable energy. You see: The WMF does not need to generate its own energy (that's the idea of a market: Some are producing one product some a different, so not everyone has to do everything on his own). KPFC💬 10:15, 10 February 2017 (UTC)[reply]

Against![edit]

It has been noted that insects can be attracted to the bright light caused by concentrated solar technology, and as a result birds that hunt them can be killed (burned) if the birds fly near the point where light is being focused. This can also affect raptors who hunt the birds.[1][2][3][4] Federal wildlife officials have begun calling these power towers "mega traps" for wildlife.[5][6][7]

Bats may be injured by direct impact with turbine blades, towers, or transmission lines. Recent research shows that bats may also be killed when suddenly passing through a low air pressure region surrounding the turbine blade tips.[8] The numbers of bats killed by existing onshore and near-shore facilities have troubled bat enthusiasts.[9]

Hi Idot, what would you propose? Thanks, --Gnom (talk) Let's make Wikipedia green! 15:26, 7 February 2017 (UTC)[reply]

Wikipedia should use only that power plants that are safe for animal (Idot (talk) 15:40, 7 February 2017 (UTC))[reply]
e.g. use PV solar power, and FIGHT AGAINST CS solar power (Idot (talk) 15:44, 7 February 2017 (UTC))[reply]
Thank you for your explaining this, Idot. I understand that you are concerned because of certain types of renewable energy. But you agree that in principle, we should stop relying on carbon fuels, correct? --Gnom (talk) Let's make Wikipedia green! 16:57, 7 February 2017 (UTC)[reply]
yes, it should be only safe renewable energy, not any renewable (it would be very bad if Wikipedia build a Huge Solar Concentrator that kills lots of animals, as it is MUCH WORSE that carbon fuels) Idot (talk) 17:43, 7 February 2017 (UTC)[reply]
Okay, Idot, thank you for this additional information. I don't think that Wikimedia should build any power plants at all, but should rather buy renewable energy from certain sources. Do you have links to any reports that I can read about how a "solar concentrator" is actually worse than carbon fuels? And how many "solar concentrators" are there in the United States? Thanks, Gnom (talk) Let's make Wikipedia green! 17:57, 7 February 2017 (UTC)[reply]
when talking about renewable energy some people like solar concentrators as they look impressive at presentations of business plans, however it the wort type of energy plant, much worse that carbon emission, so it is even better to do nothing than building huge solar concentrator (Idot (talk) 17:59, 7 February 2017 (UTC))[reply]

ARE YOU REALLY ASKING "Do you have links to any reports that I can read about how a "solar concentrator" is actually worse than carbon fuels? And how many "solar concentrators" are there in the United States?"? all links already in Wikipedia article! how you can lead Wikipedia to anything if you even don't read it? (Idot (talk) 18:02, 7 February 2017 (UTC))[reply]

it seams that you do not take really care about environment - you just want to show your name as a lead of Wikileaks project (Idot (talk) 18:04, 7 February 2017 (UTC))[reply]
I am sorry for upsetting you. I just read en:Concentrated solar power#Effect on wildlife and I understand that this is an issue. But does this mean we should continue to rely on fossil fuels? I don't think so. --Gnom (talk) Let's make Wikipedia green! 18:10, 7 February 2017 (UTC)[reply]
such denying danger of Concentrated solar by you means like you just egomaniac how do not care about environment, but just want to show you name at any cost (Idot (talk) 18:14, 7 February 2017 (UTC))[reply]
Hi Idot. Please don't attack other people like this. You can express your own opinion without being rude like that :). Jules78120 (talk) 20:13, 7 February 2017 (UTC)[reply]
Gnome deletes votes against from signing page - which is NOT FAIR! (Idot (talk) 02:28, 8 February 2017 (UTC))[reply]
Your concerns are hyperbolic and without merit. The issues raised when CSPs initially went online were addressed and solved. See One Weird Trick Prevents Bird Deaths At Solar Towers. ValarianB (talk) 17:33, 8 February 2017 (UTC)[reply]

Discussions about how much energy we use (calculations, efficiency...)[edit]

HHVM vs PHP: efficiency[edit]

Wikimedia switched from the PHP official engine to mw:HHVM some years ago. Although the blog post mainly speaks about speed, there is some efficiency gain given the mean CPU load decreased from 50% to 10%. I don’t know if there are hard figures about the power saved by the switch, but it would be very interesting regarding the GreenIT subjet. ~ Seb35 [^_^] 20:26, 10 March 2016 (UTC)[reply]

Maybe whoever processes the bills from Equinix would know whether the total power consumption decreased in 2015 compared to 2014.
It's rather easy to see average load over the last 12 months (with some disclaimers), but old data is unreliable or difficult. wikitech:Category:Servers isn't of much use and most phabricator:tag/procurement reports are unnecessarily private, but phabricator:T119598#1885991 hints WMF is on Intel Xeon E5 CPUs and standard tests on similar machines (I'm looking at the appendix) show that going from 50 % to 10 % load only decreases the power consumption by about 30 %, because maximum efficiency is reached above 50 % load. It seems that other things like blade configuration may matter more.
It would be great to know whether the WMF uses platinum-rated PSU; I asked a few years ago but RobH wasn't able to answer. Nemo 16:33, 14 March 2016 (UTC)[reply]

Buckets-O-Coal[edit]

So in the Data Centers section, there is a spreadsheet that estimates KWh and things. I think that a column should be added to better VISUALIZE what these numbers mean for those who don't respond to numbers or text. I'm thinking that we measure energy in buckets of coal or gallons of gas, etc. This site gives a decent way to estimate that: https://www.eia.gov/tools/faqs/faq.cfm?id=667&t=2 By my napkin math, if we apply the annual total 2 GWh number to the 1.04 pounds of coal from the era site and then divide by 34% (the amount of that which is coal) the total pounds of coal comes to 707,200. According to BNSF, http://www.bnsf.com/customers/equipment/coal-cars/ the carrying capacity of one freight train wagon is 121 tons, that's (121x2000) 242,000 pounds. By my estimate Wikimedia burns almost 3 full train wagons of Coal per year.

This could be a useful way to visualize the pollution we generate.

If anyone wants to show barrels of oil be my guest.

Three of these are burned every year by Wikimedia

Victorgrigas (talk) 03:09, 15 August 2016 (UTC)[reply]

What's wrong with using standard comparisons like Tonne of oil equivalent? Nemo 12:08, 15 August 2016 (UTC)[reply]
Nothing - but does it have a photo? Different people have different minds and different things appeal to them. So, If you have a numbers mind numbers might appeal, if you have a word-mind, descriptions might appeal, if a visual mind, pictures might appeal. We need different ways to describe/illustrate/communicate a measurement.Victorgrigas (talk) 12:29, 15 August 2016 (UTC)[reply]
Of course it can have a photo. It would perhaps be better to focus on figures for climate-altering emissions and show how much material (coal or other) contains as much carbon dioxide equivalent, but we can also play on the "fuel consumed" metaphor.
I don't quickly find reliable statistics on efficiency, let alone LCA or EROEI, of power plants in Texas or Virginia (the web is full of propaganda like [2]), but if we take [3] [4] ([5]) [6] and grant a generous 33 % efficiency to the average plants in Texas/Virginia (while ignoring variations in fracking, externalities and everything) we have an approximate 2 GWh / 0.33 / ( 11630 kWh/toe ) = 521 toe. This ship has a tonnage of 400 t: .
Then of course per https://xkcd.com/1257/ you can always invent other comparisons, such as (numbers invented!) "enough oil to cover the entire Atlantic ocean 1500 times" or "every pageview spills a teaspoon of oil" or similar. Nemo 13:30, 15 August 2016 (UTC)[reply]

Since we're really talking about pollution generated (rather than the amount of each fuel consumed), I propose referencing each type of energy consumed to the weight of CO2 produced by that energy source. Since there was already a "to do" note to add this carbon footprint to the existing tables, I have computed the number of tons of CO2 for each row. The total weight of CO2 generated by producing the energy used in the data centers (for example) would fill almost 8 of those largest coal cars, if the gas could somehow be squished into them. Johnson487682 (talk) 20:25, 16 January 2017 (UTC)[reply]

Calculating Wikimedia's energy use[edit]

it just says that Wikipedia is a huge website and that several things could be done to lower the carbon emission for this website. I'd say that this is pretty uncontroversial, don't you think

Really? Let's do some math:

Ganglia shows that there are 1283 servers[1], something that can be more or less checked on the puppet config[2]. Probably many of those are virtual servers, which means in reality there are much less hardware nodes in reality. But let's be conservative, and assume there are 1300 physical machines, and that way we can include other kinds of hardware, such as switches, shared storage, etc.

Average server consumption is difficult to measure, but let's assume, to simplify things, that servers are at 100% usage (it was mentioned that the average cpu usage is 9%[3], and that seems about right, because probably most of the servers are "idle", only there to provide high availability, in case the primary ones fail. Plus, for performance issues, you do not want to maximize its usage, and therefore, its energy consumption). Some references found online show that server consumption can vary between 200-1000W[4][5].

Let's take 400W.

400W * 1300 servers * 24 hours/day * 365 days per year / 1000 WH in a KWH = 4555200 KWH every year (actual measured consumption for eqiad seems to be 1000 MWH[6], so it seems the numbers are very overestimated, but within the range of plausible).

Of course, there is still a huge chunk if that that has not been acounted yet, which is energy for non-IT equipment. That is known as PUE [7]. There is not much data about this, so let's set it on an overhead of 30%* I later read that average overhead is 70%, but wikimedia servers are colocated on not-owned datacenters.

4555200 * 1.3 = 5921760 , or approximately 6 million KWH per year.

Let's compare to Facebook's own data[8]. In 2014, 985,000,000 KWH were consumed by Facebook datacenters. In comparison, 164 times more than Wikimedia's. In fact, Facebook spent 8 times more on "office and other spaces" energy consumption than all Wikimedia servers together, across all datacenters.

But Jynus, surely they have more visits than Wikipedia and friends! They do more things. It is about efficiency, not total numbers! There are no accurate stats from first parties, but Facebook was reported that it reached 1-1.5 billion active users in 2015, with a similar number of unique visitors per month.[9] [10]. With wikimedia having around 150 - 500 UVPM[11] [12][13]. That means Facebook is using 164x the energy consumption for 2x-10x the traffic (in users).

Google purchased recently green energy (payed for it, not really used directly [14]) energy by value of 1,146MW. 1600 times our consumption!

But we are doing badly? Do not get me wrong, I would support anything merely symbolic, but our figures are close to those of a household/small business, than to a factory. But I would dare anyone to tell me a similar site in impact, with Wikimedia's level of (energy) efficiency.

In my personal opinion, Greenpeace report is highly misleading when billion dollar companies, thousand of times more polluting in total carbon footprint emissions than other NGOs and small business rate better.

--Jynus (talk) 08:09, 30 March 2016 (UTC)[reply]

Hi Jynus, thank you for your comment! Even if the Wikimedia servers consume significantly less energy than those of other websites, few carbon emissions are still carbon emissions and we could avoid creating them. Plus, we could set an example for the internet sector. --Gnom (talk) 08:52, 30 March 2016 (UTC)[reply]
"few carbon emissions are still carbon emissions and we could avoid creating them." How? Like Google, with a check that actually does nothing except PR, or like Facebook creating a fancy graphic that also does nothing? Hiring a graphic designer and thinking we have done something instead of *actually* doing something is a dangerous path. And I am not against green efforts, but "awareness", depending how it is done, can be more dangerous than doing nothing- cleaning up your conscience and thinking you have done something good. I would like to see another report from another organization, having into account absolute numbers.
I believe, Wikipedia (and the rest of the Wikimedia projects) can be way more effectively impacting by creating and maintaining scientifically accurate articles about climate change en:Global_warming or informing the public about countries' political disasters such as en:Kyoto_Protocol.
--Jynus (talk) 09:58, 30 March 2016 (UTC)[reply]
Well, the measures proposed in the essay are based on Greenpeace's ideas. I suppose these people know what they're doing. Why shouldn't we use renewable energy for the server sites? --Gnom (talk) 11:10, 30 March 2016 (UTC)[reply]
Can someone confirm the numbers here? Jynus's estimate is between 1,000 mWh MWh and 4,555 mWh MWh per year, if I understand him correctly. --Gnom (talk) 08:56, 30 March 2016 (UTC)[reply]
I assume you mean MWh, not mWh. The difference is just 9 zeros. :) Nemo 12:17, 30 March 2016 (UTC)[reply]
Oh. Yes, of course. --Gnom (talk) 13:55, 30 March 2016 (UTC)[reply]

rough estimate of carbon impact of Wikimania 2017: >2x servers[edit]

I've done a very rough estimate of the carbon impact of travel to Montreal. The numbers below are very rough, but might serve to help gauge relative priorities. Assumptions made:

  1. 1500 attendees, based on recent #s from London, DC, and Esino Lario (i.e., those closer to more Wikimedians).
  2. Distribute those attendees based on my experience and some numbers from the 2014 scholarship goals.
  3. Picked large airports near the center of the relevant geographies, where possible with direct flights to Montreal.
  4. Assume all flights are economy.

Using the ICAO carbon calculator, which does not include radiative forcing in their calculation (see below), I get:

  • 700 US attendees, traveling from Dallas: 382130 kg
  • 500 EU attendees, traveling from Berlin->Toronto->Montreal: 430471 kg
  • 200 Canadian attendees, traveling from Toronto to Montreal: 25094 kg
  • 50 Central American attendees, traveling from Mexico City: 28803 kg
  • 50 Asian-Pacific Attendees, traveling from Hong Kong->Toronto->Montreal: 58410 kg
  • 25 African/Middle East attendees, traveling from Cairo: 28918 kg
  • 25 Latin American attendees, traveling from Santiago->Dallas->Montreal: 32435 kg

After applying a multiplier of 1.9x for radiative forcing, we get 1874 metric tons of C02 impact. In other words, in this very rough calculation, Wikimania is 2.25x as big a source of global warming as the servers. To put it a slightly different way, each Wikimania attendee, on average, will have the same impact as 13 hours of running the servers.

There are of course a variety of potential sources of error here:

  1. Possible sources of underestimation:
    1. Other sources give 2.7x as the correct multiplier for radiative forcing, though that appears to be based on older research.
    2. This is only Wikimania; doesn't count Wikimedia Conference, etc.
    3. Doesn't count the event itself, only travel. (Many discussions of carbon footprint of conferences include the event itself, but since the attendees would not be in hibernation during the conference, I don't think this is very useful.)
  2. Could go either way:
    1. I could not find any good demographic data on Wikimania attendees, so the location headcount estimates are very rough.
    2. The "large, central" airports are of course an approximation. Asia/Africa probably underestimate (because of extra connections); Canada probably overestimates (because many will drive or train).
    3. Wikimanias that are more distant from major centers of contribution may have fewer attendees, but longer travel times; I have not attempted to simulate that.

Given all that, the 2.25x/13 hours numbers are obviously not precise. However, I think it is safe to say that travel to events is at least as big and likely a substantially larger impact than servers.

This is not to say that we should stop caring about the servers, of course, but just to put it in perspective and stimulate consideration of options. For example, assuming we don't have clean energy options at our current colos, it might be a more effective use of Foundation money to invest heavily in video streaming and chat, and cut Wikimania attendance in half, than to move colos. And of course individual contributors who are concerned about carbon impact might consider how they can reduce their travel and increase use of on-line collaboration options. —LuisVilla (talk) 18:29, 12 February 2017 (UTC)[reply]

Hi LuisVilla, thank you for this information! We should include it on the page as well. I have many questions regarding your final paragraph, can I fire away? Thanks, --Gnom (talk) Let's make Wikipedia green! 21:23, 12 February 2017 (UTC)[reply]
@Gnom: Sure, of course fire away, and feel free to double-check the #s as well! —LuisVilla (talk) 22:09, 12 February 2017 (UTC)[reply]
Great, LuisVilla! I have added your estimate to the page. Now to my questions:
  • You wrote that your post was meant to "stimulate consideration of options". I assume that you agree that a reduction of air travel is not an "option" in the sense that it excludes using renewable energy for Wikipedia's servers. Do you see these two related in another way?
  • Also, you wrote "assuming we don't have clean energy options at our current colos [...]" Do you think this could be the case? Coming from a country where switching to renewable energy takes nothing more than a phone call to your energy provider, I am having a hard time imagining this scenario.
  • Next, should we really consider to "cut Wikimania attendance in half"? In my mind, this would mean a lot of torches and pitchforks... So maybe better start with the servers?
  • Then, you wrote that looking at air travel might be more effective than "to move colos". Could it not be possible to ask the existing colocation providers to run "our" servers on renewable energy?
  • Finally, I know for a fact that "individual contributors who are concerned about carbon impact" from Europe are already considering not attending Wikimania and buying buying carbon offsets at their own cost in order to ease their consciences (I just saw a discussion about this on Facebook the other day).
I hope these questions can help move the discussion forward. And again, thank you so much for weighing in! --Gnom (talk) Let's make Wikipedia green! 17:16, 13 February 2017 (UTC)[reply]
@Gnom: I don't follow your first question? As to the others:
  • With regards to switching, depending on the area of the US and how the colo provides electricity (since we don't run our own colos), no, WMF may not have that option. The US isn't all on a single grid, so what may work in California (where we do have that option for the office) may not work in Texas, and what may work when you're Amazon or Google and build your hosting facility from scratch may not work when you are on a shared system with many other providers. I have not looked at the details in particular colos or states, though. But again, this is not my area of expertise; as you say, it may be as easy as calling the colo and having them talk to the power provider, and it would be nice if WMF IT clarified if that was an option, and if so, why it hasn't been done.
  • I am somewhat agnostic on Wikimania: it is obviously good in many ways, and I enjoy it personally when I can go, but it is also extremely expensive even putting aside the carbon impact, and there are reasons to think it is not as effective overall as more local conferences. (Largely because of Wikimania, the Foundation spends more on travel+conferences than on "internet hosting" in 2014-2016.) It is possible that money could be spent more effectively, and perhaps CO2 is just one more reason to look seriously at those expenses and how the conference could be reconfigured.
  • Sure, I buy my own carbon offsets for all my travel (lately from Solar Aid, which also has a nice educational benefit), as I suspect others in the movement do. It would be nice if, say, WMF travel scholarships included a Solar Aid offset. But at this time that's a luxury, and as someone pointed out when discussing the servers, offsets aren't actually a solution to the greater problem.
Hope that helps. —LuisVilla (talk) 17:53, 13 February 2017 (UTC)[reply]
@LuisVilla: it would be nice if WMF IT clarified if that was an option Yes, that would in fact be very nice. If you have an idea who to talk to, let me know. I have been unsuccessful at this for quite some time now. It's been one and a half years, actually. --Gnom (talk) Let's make Wikipedia green! 23:09, 16 February 2017 (UTC)[reply]
Somehow the result of this calculation seems a little unrealistic to me. If we take a standard plane[7] the emissions are 2.5 tons per hour, which means at most ~8 tons CO2 per hour (factor 3.15)[8], assuming complete combustion. If we then assume an average flight duration of 6 hours (results from the cities listed above) and calculate a need of 1500/200≈8 full planes that gives us 8*6*8=384 tons CO2, which is significantly less than what Luis calculated above (even if including radiative forcing), despite these non-conservative estimations. I admit this calculation is even rougher, but I would like to understand where this huge difference in result could come from. --Vogone (talk) 17:34, 15 March 2017 (UTC)[reply]
@Vogone: You can double-check my calculations with the ICAO or Atmosfair calculator if you'd like. (I did not originally use Atmosfair, but checking a few of the ICAO numbers suggests they are largely in agreement.)
To look at it another way, a pretty common number (I believe originally sourced to the ICAO) suggests one NY->LAX round trip flight (about six hours) is just under 2 metric tons/passenger;[9] Atmosfair gives roughly the same 2 metric tons number for a shorter flight (Denver->NY).[10]. Your math is closer to 0.2; an order of magnitude smaller.
So I'm afraid something is wrong in your math or your sources; I'm sorry that I don't immediately see what. LuisVilla (talk) 18:04, 1 December 2017 (UTC)[reply]

Power usage graph[edit]

No idea how accurate it might be, nor what kind of power usage it measures (power needed by each machine? power actually consumed by the power supply? total power entering the cluster?), but I see https://grafana.wikimedia.org/dashboard/db/site-power-usage exists. --Nemo 20:49, 19 November 2017 (UTC)[reply]

Ah and there's some trend information at mailarchive:wikitech-l/2017-November/089058.html. --Nemo 21:29, 19 November 2017 (UTC)[reply]

Discussions about the data centers[edit]

Waste heat recycling[edit]

Another way to limit the environmental impact of servers is to use their waste heat in another building or in a heat grid, like described here. That would be really green. Any idea whether this has been proposed and/or studied?--Canaricolbleu (talk) 13:01, 4 April 2016 (UTC)[reply]

This is something the owner of the facility has to handle, of course. wmf:RFP/2013 Datacenter and wmfblog:2014/05/05/wikimedia-foundation-selects-cyrusone-in-dallas-as-new-data-center/ show there was no such consideration the last time a location provider was chosen.
Greenpeace provides some data on how well Equinix fares, I think. EvoSwitch does better but I don't remember reading independent assessments yet. Nemo 12:36, 5 April 2016 (UTC)[reply]
But we should perhaps have our own datacenters anyway. Which could be very green. Rich Farmbrough 22:03 29 April 2016 (GMT).

New datacentres[edit]

Cf. phabricator:T156029#3052763. --Nemo 11:52, 24 February 2017 (UTC)[reply]

I notice on https://obamawhitehouse.archives.gov/the-press-office/2015/11/30/white-house-announces-additional-commitments-american-business-act that «At Equinix we believe it is important that companies operate in an environmentally sustainable way. For us, electric power is our largest source of greenhouse gas emissions. As such, earlier this year, we announced a long-term commitment to use 100 percent clean and renewable energy across our global platform».
I'm curious to see how this progresses concretely. Nemo 17:01, 30 March 2017 (UTC)[reply]

Discussions about renewable electricity (definition, carbon offsets, cost...)[edit]

Re: Green power[edit]

I'm confused by the two sections on green power. Switching to green power is not complicated and does not require time. It's trivially easy to buy green certificates. It's a bit less easy to choose certificates whose seller actually adds new renewables rather than just shuffling cards with other clients. Prices in USA are a mess, of course; probably it would be a fraction of the WMF's budget, but we'd need some data to know... we currently have no idea how much power WMF consumes (or what energy class the WMF office is, or even how big perhaps). --Nemo 21:17, 1 August 2015 (UTC)[reply]

From what I understand, buying green certificates is not what Greenpeace expects us to do. They want us to actually get renewable energy for our servers, in order to avoid "shuffling cards". --Gnom (talk) 09:36, 2 August 2015 (UTC)[reply]
Any source for that? There is really no difference between buying green power and buying green certificates. It's not like you can choose "what electrons" you're delivered at home, unless you self-produce locally. --Nemo 10:03, 2 August 2015 (UTC)[reply]

Estimating the cost of renewable energy[edit]

Hi, I am now trying to estimate how much it would cost the Wikimedia Foundation to run its servers and its office on renewable energy. I have found the website of the U.S. Department of Energy's Green Power Network and I think this is where we can find out best about the feasibility and cost of this idea using the information I have gathered so far. Maybe someone can help me figure this out? Thanks, --Gnom (talk) 09:53, 30 May 2016 (UTC)[reply]

Final prices depend on time of the day, cost of delivery to the specific location etc. It makes more sense to look for the price of a green certificate or REC of suitable quality. Nemo 13:13, 31 May 2016 (UTC)[reply]
We actually know the specific location of the servers, we can estimate their energy use throughout the day and we have the clear recommendation from Greenpeace that certificates are not really an option. --Gnom (talk) 22:12, 31 May 2016 (UTC)[reply]
I said "of suitable quality", surely Greenpeace has no objection to certificates in general. Can you please point me to said recommendation, which I missed? Nemo 06:09, 1 June 2016 (UTC)[reply]
See Greenpeace USA: Clicking Clean: A Guide to Building the Green Internet, May 2015, p. 29 et seq. --Gnom (talk) 07:58, 1 June 2016 (UTC)[reply]
Which confirms what I say: «If a company executes a virtual PPA in the same electricity market in which it operates a data center, then a virtual PPA can still be a credible way to add renewable energy and displace demand for dirty energy on the same grid. [...] If companies do buy unbundled REC/GOOs, they should at a bare minimum buy RECs that demonstrate strong additionality, and are in close proximity to the facilities they wish to claim are renewable». Let me know if you need additional information on what constitutes "suitable quality". Nemo 08:27, 1 June 2016 (UTC)[reply]
From what I have learned, "renewable energy" in the proper sense must be additional, i.e. come from renewable energy sources that wouldn't otherwise be there, and local, i.e. displacing fossil fuel energy that is powering the same energy grid running the servers and offices. This is why I would like to first try to actually have the servers actually run on renewable energy. If this is not possible, then we can discuss any alternatives. Thanks, --Gnom (talk) 11:51, 1 June 2016 (UTC)[reply]
Why are you rephrasing what I just quoted? :) I agree with your goal, the best method to achieve that are certificates. Nemo 13:34, 1 June 2016 (UTC)[reply]
I'm sorry, but I understand that "buying renewable energy certificates" is not as good as actually "buying renewable energy". Am I mistaken here? Thanks, --Gnom (talk) 20:29, 1 June 2016 (UTC)[reply]
Yes, you are mistaken. Nemo 12:16, 15 August 2016 (UTC)[reply]

Greenpeace...you serious?[edit]

I agree with this initiative in principle, however I note some seriously disturbing politicization of this issue. For starters, why are we using the anti-science organization Greenpeace, as the benchmark of what is "clean energy"? We should use the IPCCs data on the life cycle greenhouse gas emissions of different energy sources, to determine how well we are doing. Not a "scorecard" from an avowed anti-science organization, with a major antipathy towards nuclear energy. Moreover, if for example, we burned down a rainforest of trees and called it "biomass energy", then this tragically would get classified as "clean energy" according to Greenpeace. With that, I hope you're beginning to see some of the problem I see, with this "scorecard" malarkey. File:Greenpeace USA analysis of Wikipedia enegery consumption 2015.JPG --Boundarylayer (talk)

This assessment doesn't seem to align much with reality. --Nemo 16:05, 1 November 2017 (UTC)[reply]
I know Nemo, Greenpeace are not a scientific nor authoritative body by any farcical stretch of the imagination, so exactly why are we promoting their analysis and clasification scheme for what is "clean energy"? Why not use the UNFCCC & IPCC accredited "life cycle greenhouse gas emissions of different energy sources"? Why are we not using actual peer-reviewed science as our yardstick for what is "clean energy"? Boundarylayer (talk) 19:55, 17 March 2018 (UTC)[reply]
Hi Boundarylayer, I think it would be best if you could remain calm and point us to a better source. Thank you, --Gnom (talk) Let's make Wikipedia green! 20:15, 17 March 2018 (UTC)[reply]

Not-for-profit electric power preference by Wikimedia Foundation?[edit]

Gnom- Can the Wikimedia Foundation vote to support local publicly-owned, non-profit electric power agencies, municipalities and cities world-wide?

What does this even mean? There are some cooperatives for renewable power generation and self-consumption, is this what you're talking about? --Nemo 12:27, 14 September 2017 (UTC)[reply]

Discussions about the offices[edit]

Impact[edit]

There's a ton of things that any bricks-and-mortar organisation can do to reduce impact:

  • Use quality hot-air dryers instead of paper towels
  • Avoid disposable cups
  • Xeriscaping for any horticultural areas
  • Using zero water urinals
  • Encouraging remote working and virtual meetings
  • Setting limits for use of heating and air con.
  • Minimising paper use.
  • Encouraging cycling, car sharing.
  • Using LED lighting.
  • Use powdered or tinned milk instead of having an office fridge.

Rich Farmbrough 02:47 1 May 2016 (GMT).

Some more suggestions:

  • Employ fewer people -- use less office space, fewer workstations
  • Buy only vegetable foodstuffs and only from local suppliers
  • Locate offices in more temperate climates, cities with sustainable housing and transport
  • Locate data centres in locations with geo or hydro power

Rogol Domedonfors (talk) 15:53, 3 June 2016 (UTC)[reply]

I think we should see more global, the foundation should compensate all the pollution generated by the Wikipedians, because it must represent quite a lot of emissions. Not everyone can travel by bike and train as I do. It would be necessary to see if the foundation could not build some wind turbines and some solar parks to make this compensation. The cost of renewable energy is absolutely not important, it is about showing the right way and being exemplary. Jérémy-Günther-Heinz Jähnick (talk) 18:51, 5 February 2017 (UTC)[reply]

Suistanability & Efficiency[edit]

If you want to reduce carbon- or fuel burning (& Atomic-Energy-Waste too), so It will help you to understand the Differences by the Efficiencies, the way "How" you transform Energy into Heat or Electricity.

So if you want to reduce anything, you have to start at the Beginning. The Beginning of the Process is the Transformation - here you can save a lot if you exactly know how many Energy you can transform per Unit. Most of the Heating-Systems we use have much less Results, than we be suggested or anybody believes. Look at the "German Example" - it seems destructive only prefer the Insulation-Systems they were called "WDVS-System". Some People are confused about the Results and are thinking about loud.

The better the outer walls are insulated, the worse the efficiency of the heating system becomes. This is significantly negative and does not meet the desire to reduce energy consumption.

See the technical term "Anlagenaufwandszahl ep", which is used in Germany to assess the annual utilization rate.

Anlagenaufwandszahlen ep, Jahresnutzungsgrade. Wirkungsgrade, Energietransformation. Energienutzung, Effizienz, Suisanables, Umweltheizung
Results of Efficiency after Insulation like EnEV

The EnEV is a Order by the Gouvernement, they published own fiscal Interests and have to work for their Benefits, the Economies have paid for. This is not real - it is a Nightmare in Politics.

User: Umweltheizung, signed in and written at 16.01.2018, 18.47 MEZ

Frage: Is it allowed to place the .png-file Anlagenaufwandszahlen < 1,0 directly nearby the other picture which is on the right side placed?

I do not have the Permission User: Umweltheizung/Vandalism

Food and materials[edit]

As a professional chef, I know well the need for sustainability beyond clean energy. For the WMF's ~280 employees (not counting any on-site volunteers), where is their food coming from? Is there a recycling system in place? Are disposable plates, cups, or utensils used, or something more environmentally friendly (e.g. ceramic, or compostable single-use items)? Can't they set up a composting program? San Francisco surely has many outlets for that. There's more elements here to address, but this is a start. Food waste is a ridiculously important issue, and I could provide much more material on why. (talk) 00:22, 12 January 2018 (UTC)[reply]

Discussions about events[edit]

Wikimania local transport[edit]

Above, the #rough estimate of carbon impact of Wikimania 2017: >2x servers did not count the event itself, only travel. Although travel by air has huge environmental impact, we should consider also possibilities to reduce impact by local arrangements, and make such choices explicit. One reason is "We have a powerful brand and can use it to show green leadership" as stated high up on the Sustainability Initiative page.

There are now no recommendations on the Wikimania Travel tips page. For those coming e.g. from New York there are options to avoid private cars. I think there should be a discussion (or at least a note and a link to a discussion) before the "By" sections, which more or less suppose you have already chosen your means of transport.

It is nice that the event is in a city with good public transport, and probably quite high awareness about environmental issues. That alone, and even more if made obvious, might have an impact on people coming from places where awareness is lower and environment friendly arrangements less common. Something to think about also when choosing future locations and venues.

The Local transport page seems to cover public transport and bike rental as main options, good (but taxi seems to still be treated as such, why?). Private cars are not mentioned, if that means "forget about your car for getting to the conference centre" it could be mentioned explicitly. I hope the section will be expanded before the event, to make using environment friendly options as easy as possible, especially for those not used to them.

--LPfi (talk) 08:58, 3 April 2017 (UTC)[reply]

Discussions about investments[edit]

Investments[edit]

More on investments: https://www.weforum.org/whitepapers/renewable-infrastructure-investment-handbook-a-guide-for-institutional-investors/ (especially relevant since wmf:Resolution:Establishment of Endowment and wmf:Wikimedia Foundation Investment Policy). --Nemo 15:01, 3 January 2017 (UTC)[reply]

The news on the topic are frequent but often overdramatic, see e.g. [11]. Better focus on scientific or at least data-based studies. --Nemo 12:29, 14 September 2017 (UTC)[reply]

Discussions about governance (sustainability resolution...)[edit]

The case for ESG[edit]

I don't see anyone here actually making a case for changing WMF's policies. So I thought I would have a go at that.

The question is whether it is appropriate to spend donor money on an activity which, while apparently ethical, is not included in our mission statement and was not put to donors during fundraising. Broadly, should ESG or "triple bottom line" considerations should apply to charities such as ours?

In the for-profit world, the ESG movement was born out of a desire to attract responsible investors. It is partly PR driven: bad press in response to terrible events like the Exxon Valdez oil spill leads to negative screening by fund managers. Reducing the risk of bad PR reduces the risk of capital flight.

The argument for ESG considerations in a charity supported by small donors is essentially identical. Bad PR could put off donors. Responsible major donors may avoid donating to Wikimedia on the basis of environmental policies (or lack thereof) even without issues being raised in the media.

Because we are a charity and we believe we are doing good in the world, we might believe that we are somewhat insulated from bad PR. But Wikimedia has never had the full support of the press, there have always been negative stories. And the "D" rating Greenpeace gave to Wikimedia is an indication that environmental charities are not afraid to criticise non-environmental charities.

Perhaps Wikimedia's activities should be narrowly constrained by its mission statement. But is greenhouse gas abatement really an activity? I would argue that greenhouse gas emission is an activity which we have entered into as part of the cost of doing business -- like hiring employees or renting property. When we perform an activity, we have a social responsibility to perform it ethically. When we hire employees, we try to give them good working conditions. I think buying power carries with it a responsibility to mitigate the environmental damage caused by power generation.

-- Tim Starling (WMF) (talk) 04:11, 6 April 2016 (UTC)[reply]

Hi Tim, thank you for your thoughts! You are absolutely right by pointing to the issue that the environmental impact of the Wikimedia movement is a topic that seems to have no easy-to-establish direct link to our mission. A few months ago, I tried to come up with such a reasoning:
"The Wikimedia Foundation's mission is to "to empower and engage people around the world to collect and develop educational content under a free license or in the public domain, and to disseminate it effectively and globally." However ideal this goal may sound, everything comes at a cost. One such cost is the environmental impact of running the Wikimedia projects and the Wikimedia Foundation as an organization. Being part of a volunteer-based movement which aims at spreading knowledge globally and thereby improving life for every person on the planet, it is not far-fetched for the Wikimedia Foundation to think about the sustainability of its actions and to make a strong commitment in that direction.
I'd love to hear more comments on this. When I discuss the topic with non-Wikipedian friends here in Germany, most of them react with something like, "Huh, I'd have thought the Wikipedia servers would surely run on renewable energy." So at least for them, spending donor money on this is a no-brainer. Thanks again, --Gnom (talk) 06:51, 6 April 2016 (UTC)[reply]
I think that is right. In the same way that WMF needs to be a responsible employer (if it is to be an employer...), and a law-abiding corporate citizen.
There is a valid point though, the primary responsibility is to deliver the Foundations goals, albeit not at any cost.
Rich Farmbrough 22:11 29 April 2016 (GMT).

WMF board resolution[edit]

Big news: The Wikimedia Foundation Board of Trustees has adopted a resolution stating that the Wikimedia Foundation is committed to seeking ways to reduce the impact of its activities on the environment. --Gnom (talk) Let's make Wikipedia green! 13:28, 29 March 2017 (UTC)[reply]

Good news. Let's hope it will not be only words :). Jules78120 (talk) 16:47, 29 March 2017 (UTC)[reply]
There's some movement at Wikimedia Endowment. Looks like there's much to do on the investment side. Good to have some basic information about it. --Nemo 12:28, 14 September 2017 (UTC)[reply]

Discussions about this initiative[edit]

A Logo?[edit]

Hi, I've thought about creating a logo for this project – and I've already created two sketches. What do you think? Should there be a logo? What should it look like? Thanks, --Gnom (talk) Let's make Wikipedia green! 16:54, 11 February 2017 (UTC)[reply]

I like the first design personally (and there could be interesting variants); I don't think it necessarily prevents from also having a small text slogan as part of it. 76.10.128.192 21:02, 25 February 2017 (UTC)[reply]
Okay! Maybe "make Wikipedia sustainable?" Thanks, --Gnom (talk) Let's make Wikipedia green! 22:01, 25 February 2017 (UTC)[reply]
Added another idea. --Gnom (talk) Let's make Wikipedia green! 22:27, 28 February 2017 (UTC)[reply]
So I asked around on Twitter and someone greatly improved on idea C. --Gnom (talk) Let's make Wikipedia green! 11:25, 11 March 2017 (UTC)[reply]

Move page[edit]

Hi, I would like to move this page to Sustainability Initiative. Please let me know if you have any concerns. Thanks, --Gnom (talk) Let's make Wikipedia green! 22:07, 14 March 2017 (UTC)[reply]

I've performed the move, but feel free to voice if you disagree it can always be undone. --Base (talk) 00:54, 15 March 2017 (UTC)[reply]
We use sentence case, that uppercase "I" is unwarranted. Nemo 22:41, 15 March 2017 (UTC)[reply]
Is this any different from e.g. Wikimedia Conference? --Base (talk) 23:51, 15 March 2017 (UTC)[reply]

English wikipedia userbox[edit]

I've created a userbox for users on the English wikipedia en:User:Salix alba/Sustainability Initiative Userbox. --Salix alba (talk) 19:33, 8 October 2017 (UTC)[reply]

Babel in dewiki[edit]

Liebe Wikipedianer,

für die Benutzerseiten der deutschsprachigen Wikipedia habe ich eine Babel-Vorlage erstellt, mit dem sich die Unterstützung für die Nachhaltigkeitsinitiative ausdrücken lässt: Siehe hier.

Dear Wikipedians,

if you have a user page in the German Wikipedia, you may now use a Babel template expressing your support of the Sustainability Initiative: See template.

Beste Grüße Best regardsRübenkopf (talk) 08:23, 11 April 2017 (UTC)[reply]

Presenting the Sustainability Initiative at Wikimania[edit]

Hi, I have created a submission to talk about the Sustainability Initiative at the 2017 Wikimania. Please take a look :-) --Gnom (talk) Let's make Wikipedia green! 12:43, 26 March 2017 (UTC)[reply]

Congrats![edit]

Gnom, I congratulate you, as this project is an amazing one. I was genuinely happy when I knew of the success of the project and of the resolution. Keep going! SammyMajed (talk) 11:41, 9 April 2017 (UTC)[reply]

Who is a contact for this?[edit]

The page lists a couple of "* (WMF)" accounts but those users have no page here nor are they accessible by email. I'd like to get in touch with someone who makes decisions about alternative energy for the WMF. —Justin (koavf)TCM 05:32, 10 April 2017 (UTC)[reply]

Hi Koavf, I am currently in discussions with Jaime Villagomez (the WMF CFO) and Victoria Coleman (the WMF CTO) about this. Would you like to be added to these conversations? Regards, --Gnom (talk) Let's make Wikipedia green! 09:22, 10 April 2017 (UTC)[reply]
@Gnom: Very much so. My email address is public but I don't want to write it in plain text here in case of spambots. —Justin (koavf)TCM 16:15, 10 April 2017 (UTC)[reply]

"ROAD MAP?" for renewable energy[edit]

- where is this map posted?A ri gi bod (talk) 14:32, 30 June 2017 (UTC)[reply]

Signpost article[edit]

Hi, I wrote an article for the Wikipedia Signpost, which is going to be published soon. Please let me know your thoughts! --Gnom (talk) Let's make Wikipedia green! 10:46, 23 September 2017 (UTC)[reply]