The emissions from travel it took to report this story were 0kg CO2. The digital emissions from this story are an estimated 1.2g to 3.6g CO2 per page view.
Maybe this is what finally could push companies towards making lighter and more efficient websites? At least an estimate of a webpages watt usage (server and client side).
I'm having a very hard time rationalizing these numbers. Something has to have some very poor assumptions:
Let's assume each page load is ~1mbit, and it's being served from 1000km away. 1mbit may be low, but most of it will presumably be content delivered by a CDN which is far closer than 1000km.
The reference design (which I work on) for 400G DWDM systems are ~100W for 2 transceivers + EDFA + switches, for a standard 80km link. Assuming we hop every 80km (which should be conservative here), we get 3mJ for the energy through the network. The previous generation is probably 10x worse, so ~30mJ.
Let's add the endpoint too; Wifi router + cable modem (12W or rather 3W marginal power since it's running anyway/ 200mbit) = 15mJ
And finally for the server. I'd hope the server could do 1k requests per second, and that should amortize against backends, etc (2k RPS, but frontend + DB backend server, maybe), maybe 150W for the server, so 150mJ.
This gives us 200mJ for a page load. Using California's current (1:28am) emissions for power generation of 0.274 mT/MWh (it's 1/3rd of this during the day), this yields 15ug/request.
Where do the remaining 5 orders of magnitude come from? What am I missing? If TCO/production energy of the server was included, that couldn't more than double it (capex in a datacenter < opex, and if 100% of the server cost was for energy, we could double, but this is wildly conservative)
I think most of it comes from your client device. Your laptop may be using 50W of power or about 20g of CO2 per hour so reading the article for 6 minutes is ~2g of CO2.
That makes quite a lot of sense, but it also makes me frown a bit at the phrasing, which I feel is quite misleading.
"The digital emissions from this story are an estimated 1.2g to 3.6g CO2 per page view."
Re-reading it, technically they are right, if "view" is taken in a non-technical fashion. But everyone reading that text would assume if they click refresh, the emissions are doubled.
Also, I do wonder if they take into account bounced users - I would not be surprised if most people only spent 5 or so seconds on the page before moving on.
I'd probably use closer to 10W as the device power level (a standard laptop, what I'm using right now), so the number for me would get close to 0.2g, but your point stands; the article is answering a different question than I was assuming, but regardless it seems that the majority of the energy is clearly the local device.
Definitely the peak would be higher. My laptop has a ~60Wh battery and gets ~6hours of battery life, so I was just assuming 10W average for web browsing.
Rest assured that you can make the numbers even worse by using a VPN thus requiring even more intermediate servers and cryptographic computations.
But before admonishing users please consider defending Net Neutrality and copyright reforms so that theses workaround become useless (for general usage). And while you are at it don't refrain from taxing ad-networks that use bandwidth and cpu cycle with over-engineered adware!
But nah, better stick to boomers worldviews, Governments and Big Corp can't do nothing for the planet because people are so dumb ¯\_(ツ)_/¯
It's the bbc - accuracy is the very least of their concerns. They're much more focused on harassing pensioners for their state sponsored extortion racket money.
The government continues to reduce the amount of money they send to the BBC, which continues to produce world-class television and is struggling to find the money to do so.
They've decided that they will no longer offer the complimentary license to the elderly.
None of that should have any bearing on the accuracy of their environmental reporting.
The average human breathes out 1g of CO2 every minute and a half.
By comparison, average UK cars put out 124.5g for every km driven.
It's not nothing, but it's not a lot either. It probably is worth evaluating the electrical efficiency of the tech industry as a whole, both in terms of highly executed code hotspots and entire pieces of business of questionable utility (ad javascript, cryptocurrency proof-of-waste), but individual web pages are the plastic straws of the industry compared to the giant flare stacks of real problems.
So it takes the computing power equivalent of a 300W PC running for 10 hours to serve a thousand page views? At 50 req/page view, that server is handling ~1.5 requests per second.
That kind of throughput can be handled by a RaspberryPi, so are we seriously wasting 95%+ of the power budget routing packets and idling servers?
I remember writing a paper in college about what the environmental impact of a web page was. One of the biggest energy savings Identified by far was changing the color of your website to use dark backgrounds.
Comments
Maybe this is what finally could push companies towards making lighter and more efficient websites? At least an estimate of a webpages watt usage (server and client side).
Holy cow, that is a lot. 1-3 grams of CO2 per single load of a page? Makes me feel guilty about even clicking the article.
I'm having a very hard time rationalizing these numbers. Something has to have some very poor assumptions:
Let's assume each page load is ~1mbit, and it's being served from 1000km away. 1mbit may be low, but most of it will presumably be content delivered by a CDN which is far closer than 1000km.
The reference design (which I work on) for 400G DWDM systems are ~100W for 2 transceivers + EDFA + switches, for a standard 80km link. Assuming we hop every 80km (which should be conservative here), we get 3mJ for the energy through the network. The previous generation is probably 10x worse, so ~30mJ.
Let's add the endpoint too; Wifi router + cable modem (12W or rather 3W marginal power since it's running anyway/ 200mbit) = 15mJ
And finally for the server. I'd hope the server could do 1k requests per second, and that should amortize against backends, etc (2k RPS, but frontend + DB backend server, maybe), maybe 150W for the server, so 150mJ.
This gives us 200mJ for a page load. Using California's current (1:28am) emissions for power generation of 0.274 mT/MWh (it's 1/3rd of this during the day), this yields 15ug/request.
Where do the remaining 5 orders of magnitude come from? What am I missing? If TCO/production energy of the server was included, that couldn't more than double it (capex in a datacenter < opex, and if 100% of the server cost was for energy, we could double, but this is wildly conservative)
Here is the methodology https://www.bbc.com/future/article/20200131-why-and-how-does...
I think most of it comes from your client device. Your laptop may be using 50W of power or about 20g of CO2 per hour so reading the article for 6 minutes is ~2g of CO2.
That makes quite a lot of sense, but it also makes me frown a bit at the phrasing, which I feel is quite misleading.
"The digital emissions from this story are an estimated 1.2g to 3.6g CO2 per page view."
Re-reading it, technically they are right, if "view" is taken in a non-technical fashion. But everyone reading that text would assume if they click refresh, the emissions are doubled.
Also, I do wonder if they take into account bounced users - I would not be surprised if most people only spent 5 or so seconds on the page before moving on.
Also why not add sustinance cost for the human looking at the device then? Food, water, heating/cooling, clothes, glasses, etc...
Proportionally for 2 minutes of course.
Those costs would be the same if the article was read from a newspaper or magazine, so they can be factored out.
My desktop is on whether I'm reading the article or not so those costs can be factored out too.
Surely "view" should also include the human doing the viewing? :P
According to https://news.ycombinator.com/item?id=24068986, that's 1g every 90s, so 4g CO2 for 6 minutes reading time.
In case someone wants to do the math again :
50W for 1 hour : 50Wh or 0.050kWh
California : 187gCO₂eq/kWh according to https://www.electricitymap.org/zone/US-CAL-CISO
-> 9.35gCO₂eq/h -> 0.16g/min -> 0.935g for 6 minutes
I'd probably use closer to 10W as the device power level (a standard laptop, what I'm using right now), so the number for me would get close to 0.2g, but your point stands; the article is answering a different question than I was assuming, but regardless it seems that the majority of the energy is clearly the local device.
Isn't a modern laptop mostly idle after loading the page (< 5 seconds) ?
There would be a peak for the first seconds, then less than 10W for the rest of the reading time ?
Definitely the peak would be higher. My laptop has a ~60Wh battery and gets ~6hours of battery life, so I was just assuming 10W average for web browsing.
Thanks for the figures. I've bought an ARM PC in part for lowering its footprint.
Rest assured that you can make the numbers even worse by using a VPN thus requiring even more intermediate servers and cryptographic computations.
But before admonishing users please consider defending Net Neutrality and copyright reforms so that theses workaround become useless (for general usage). And while you are at it don't refrain from taxing ad-networks that use bandwidth and cpu cycle with over-engineered adware!
But nah, better stick to boomers worldviews, Governments and Big Corp can't do nothing for the planet because people are so dumb ¯\_(ツ)_/¯
How is this comment admonishing users? It looks like an attempt to put mathematical justification to a claim.
The comment is not admonishing anyone, it's about the underlying narrative the comment is addressing.
Sorry my ironic and rhetorical intentions might not have been clear enough, English is not my mother tongue so I might been confusing here.
Basically my point is that heavy calculation about "X is emitting Y shame on you people" is too often used as decoy so you never actually audit Z.
It's the bbc - accuracy is the very least of their concerns. They're much more focused on harassing pensioners for their state sponsored extortion racket money.
I'm not sure that's relevant to this discussion.
The government continues to reduce the amount of money they send to the BBC, which continues to produce world-class television and is struggling to find the money to do so.
They've decided that they will no longer offer the complimentary license to the elderly.
None of that should have any bearing on the accuracy of their environmental reporting.
The average human breathes out 1g of CO2 every minute and a half.
By comparison, average UK cars put out 124.5g for every km driven.
It's not nothing, but it's not a lot either. It probably is worth evaluating the electrical efficiency of the tech industry as a whole, both in terms of highly executed code hotspots and entire pieces of business of questionable utility (ad javascript, cryptocurrency proof-of-waste), but individual web pages are the plastic straws of the industry compared to the giant flare stacks of real problems.
It's the equivalent of driving 19 metres in an average UK car. That's not far, but that is for a single web page. That seems incredibly high to me.
What is the comparison of the number of cars in the world versus the number of people in the world ?
Estimates of vehicles in use (including commercial) seem to be about 1.2 billion. But the relevant number is surely total miles driven?
Ads and all the services that make a modern Web page should be included in the estimation.
That is not a lot. it's about 4mWh for coal power.
In comparison, leaving a 300W computer setup on for 10 hours each day with coal power is 3KWh or 2.5kg of CO2.
So it takes the computing power equivalent of a 300W PC running for 10 hours to serve a thousand page views? At 50 req/page view, that server is handling ~1.5 requests per second.
That kind of throughput can be handled by a RaspberryPi, so are we seriously wasting 95%+ of the power budget routing packets and idling servers?
A million, not a thousand. Milliwatt-hours to kilowatt-hours.
Sending a short email is estimated to release 4g CO2:
https://phys.org/news/2015-11-carbon-footprint-email.html
When I reached the bottom of the page it even loaded an extra article. Now that's a real waste of CO2.
Just watching a movie about CO2 emissions is order of magnitudes above.
Not the whole story. The link provides more details (https://www.bbc.com/future/article/20200131-why-and-how-does...)
It involves counting carbon from journalists' travel etc. So a full tally (which is good but isn't just about literal pages loaded)
I also wonder how much JS + adverts have to offer as part of that carbon load.
We need an open source model for estimating the CO2 emissions of a server, of the network (including CDNs, etc.), of a device.
I was also disappointed with the amount of text compared to the amount of information in the article.
It seemed like every paragraph started with for sure you have been in the waves so you know they are waves
Agreed, far too much waffle. The first three paragraphs don't even mention the new ship design.
Why on earth would it finally push companies to do anything? Where is the profit motive in this?
I am not against lighter websites. But it's not gonna happen unless we start pricing carbon.
I recently saw another site that tries to offset its carbon footprint: https://mynoise.net/carbonFootprintCompensation.php
I hope this becomes more prevalent at least for paid services.
I remember writing a paper in college about what the environmental impact of a web page was. One of the biggest energy savings Identified by far was changing the color of your website to use dark backgrounds.