4 ms·

Saving CO2 using location and time shifting in Azure

34 points by komape 3y ago

toomuchtodo 3y ago
Love the ElectricityMaps shoutout [1]. For those who may say, "why?", consider orgs that are climate/carbon aware as well as upcoming supply chain emissions reporting obligations (US SEC [2] [3] [4] [5]). If you can shift your workloads to lower carbon intensity time slices, please consider doing so! [1] https://app.electricitymaps.com/map?wind=false&solar=false https://app.electricitymaps.com/map?wind=false&solar=false [2] https://www.sec.gov/news/press-release/2022-46 https://www.sec.gov/news/press-release/2022-46 [3] https://www.sec.gov/securities-topics/climate-esg https://www.sec.gov/securities-topics/climate-esg [4] https://www.wsj.com/articles/secs-climate-disclosure-rule-isnt-here-but-it-may-as-well-be-many-businesses-say-854789bd https://www.wsj.com/articles/secs-climate-disclosure-rule-is... [5] https://www2.deloitte.com/us/en/pages/audit/articles/sec-climate-disclosure-guidance.html https://www2.deloitte.com/us/en/pages/audit/articles/sec-cli... (no affiliation with any org mentioned)
infecto 3y ago
Neat use of data. Probably not possible for Azure to do this but the way to drive this kind of change would be dynamic pricing of compute based on the localized generation cost. Lots of road blocks to implement this including quantifying how the DC is getting charged but I think interesting and the most probable way to incentive companies to do this. In general ESG reporting is pretty lackluster since its hard to set the acceptable metrics.
- mkesper 3y ago
  CO² tax would likely be effective here.
ezfe 3y ago
Reminds me of iPhone Clean Energy Charging - making small adjustments to charging patterns across all iPhones to have a (hopefully) measurable impact
- forward1 3y ago
  Is the impact actually even measurable, let alone meaningful? Or is it simply green-washing to sell more product to a consumer base primed for such marketing?
  - anoxor 3y ago
    [flagged]
    - dvko 3y ago
      Not this again. This is a number that has been doing the rounds but can be dismissed as false after less than 30 seconds of using Google and a tiny bit of brainpower. That I had to read this on Hacker News saddens me. It’s at least 33%, not 3%. Also it’s the total amount of CO2 in the atmosphere that matters, not so much the proportion. Pretty much all CO2 above 1850 baseline levels are the direct result of human-caused emissions.
      - fwungy 3y ago
        [flagged]
        anoxor 3y ago
        Add to that the fact that 30% of physics and chemistry can't be replicated. Science is not an unbiased and altruistic endeavor these days. It's not surprising the sugar industry funded science demonizing fat, cigarette companies "proved" that they didn't cause cancer, etc
        Two4 3y ago
        Of course there's ideology on all sides - almost everything we do is informed by one or more ideologies. The real question is which ideology is correct, or at least more correct? There's one side, which seems to rely on tenuous arguments based on poor interpretations of select evidence, or the other, which is based on a massive expert consensus centered on a mountain of evidence which all seems to lend itself to a single general conclusion. And then there's people like you, whose refusal to meaningfully engage is very convenient for continuing the status quo, a position that suits the first side just fine.
wizerno 3y ago
The link 404s and should be updated to https://devblogs.microsoft.com/ise/saving-co2-using-location-and-time-shifting-in-azure https://devblogs.microsoft.com/ise/saving-co2-using-location... instead

34 points by komape 3y ago

Love the ElectricityMaps shoutout [1]. For those who may say, "why?", consider orgs that are climate/carbon aware as well as upcoming supply chain emissions reporting obligations (US SEC [2] [3] [4] [5]). If you can shift your workloads to lower carbon intensity time slices, please consider doing so! [1] https://app.electricitymaps.com/map?wind=false&solar=false https://app.electricitymaps.com/map?wind=false&solar=false [2] https://www.sec.gov/news/press-release/2022-46 https://www.sec.gov/news/press-release/2022-46 [3] https://www.sec.gov/securities-topics/climate-esg https://www.sec.gov/securities-topics/climate-esg [4] https://www.wsj.com/articles/secs-climate-disclosure-rule-isnt-here-but-it-may-as-well-be-many-businesses-say-854789bd https://www.wsj.com/articles/secs-climate-disclosure-rule-is... [5] https://www2.deloitte.com/us/en/pages/audit/articles/sec-climate-disclosure-guidance.html https://www2.deloitte.com/us/en/pages/audit/articles/sec-cli... (no affiliation with any org mentioned)

infecto 3y ago

Neat use of data. Probably not possible for Azure to do this but the way to drive this kind of change would be dynamic pricing of compute based on the localized generation cost. Lots of road blocks to implement this including quantifying how the DC is getting charged but I think interesting and the most probable way to incentive companies to do this. In general ESG reporting is pretty lackluster since its hard to set the acceptable metrics.

mkesper 3y ago
CO² tax would likely be effective here.

ezfe 3y ago

Reminds me of iPhone Clean Energy Charging - making small adjustments to charging patterns across all iPhones to have a (hopefully) measurable impact

forward1 3y ago
Is the impact actually even measurable, let alone meaningful? Or is it simply green-washing to sell more product to a consumer base primed for such marketing?
- anoxor 3y ago
  [flagged]
  - dvko 3y ago
    Not this again. This is a number that has been doing the rounds but can be dismissed as false after less than 30 seconds of using Google and a tiny bit of brainpower. That I had to read this on Hacker News saddens me. It’s at least 33%, not 3%. Also it’s the total amount of CO2 in the atmosphere that matters, not so much the proportion. Pretty much all CO2 above 1850 baseline levels are the direct result of human-caused emissions.
    - fwungy 3y ago
      [flagged]
      - anoxor 3y ago
        Add to that the fact that 30% of physics and chemistry can't be replicated. Science is not an unbiased and altruistic endeavor these days. It's not surprising the sugar industry funded science demonizing fat, cigarette companies "proved" that they didn't cause cancer, etc
      - Two4 3y ago
        Of course there's ideology on all sides - almost everything we do is informed by one or more ideologies. The real question is which ideology is correct, or at least more correct? There's one side, which seems to rely on tenuous arguments based on poor interpretations of select evidence, or the other, which is based on a massive expert consensus centered on a mountain of evidence which all seems to lend itself to a single general conclusion. And then there's people like you, whose refusal to meaningfully engage is very convenient for continuing the status quo, a position that suits the first side just fine.

wizerno 3y ago

The link 404s and should be updated to https://devblogs.microsoft.com/ise/saving-co2-using-location-and-time-shifting-in-azure https://devblogs.microsoft.com/ise/saving-co2-using-location... instead