Skip to content

Latest commit

 

History

History
178 lines (137 loc) · 5.33 KB

README.md

File metadata and controls

178 lines (137 loc) · 5.33 KB
Raw

Cloud Carbon Footprint

Note

CCF is a community model, not part of the IF standard library. This means the IF core team are not closely monitoring these models to keep them up to date. You should do your own research before implementing them!

"Cloud Carbon Footprint is an open source tool that provides visibility and tooling to measure, monitor and reduce your cloud carbon emissions. We use best practice methodologies to convert cloud utilization into estimated energy usage and carbon emissions, producing metrics and carbon savings estimates that can be shared with employees, investors, and other stakeholders." - CCF

Parameters

model config

  • vendor: the cloud platform provider, e.g. aws
  • instance-type: the name of the specific instance being used, e.g. m5n.large

observations

  • cpu-util: percentage CPU utilization for a given observation
  • duration: the amount of time the observation covers, in seconds
  • timestamp: a timestamp for the observation

Returns

  • embodied-carbon: carbon emitted in manufacturing the device, in gCO2eq
  • energy: energy used by CPU in kWh

IEF Implementation

IEF reimplements the Cloud Carbon Footprint methodology from scratch conforming to the IEF specification. This means the CCF models can be run inside IEF without any external API calls and can be invoked as part of a model pipeline defined in an impl.

Cloud Carbon Footprint includes calculations for three cloud vendors: AWS, Azure and GCP.

The general methodology is as follows:

Total CO2e = operational emissions + embodied Emissions

Where:

Operational emissions = (Cloud vendor service usage) x (Cloud energy conversion factors [kWh]) x (Cloud vendor Power Usage Effectiveness (PUE)) x (grid emissions factors [metric tons CO2e])

And:

Embodied Emissions = estimated metric tons CO2e emissions from the manufacturing of datacenter servers, for compute usage

You can read a detailed explanation ofn the calculations in the CCF docs and see the code for our implementation in this repository.

Usage

In IEF, the model is called from an impl. An impl is a .yaml file that contains configuration metadata and usage inputs. This is interpreted by the command line tool, impact-engine. There, the model's configure method is called first. The model config should define a vendor and instance-type. Each input is expected to contain duration,cpu-util and timestamp fields.

You can see example Typescript invocations for each vendor below:

AWS

import {CloudCarbonFootprint} from '@grnsft/if-unofficial-models';

const ccf = new CloudCarbonFootprint();
ccf.configure({
  vendor: 'aws',
  instance_type: 'm5n.large'
})
const results = ccf.execute([
  {
    duration: 3600, // duration institute
    'cpu-util': 10, // CPU usage as a percentage
    timestamp: '2021-01-01T00:00:00Z', // ISO8601 / RFC3339 timestamp
  }
])

Azure

import {CloudCarbonFootprint} from '@grnsft/if-unofficial-models';

const ccf = new CloudCarbonFootprint();
ccf.configure({
  vendor: 'azure',
  instance_type: 'D4 v4'
})
const results = ccf.execute([
  {
    duration: 3600, // duration institute
    'cpu-util': 10, // CPU usage as a percentage
    timestamp: '2021-01-01T00:00:00Z', // ISO8601 / RFC3339 timestamp
  }
])

GCP

import {CloudCarbonFootprint} from '@grnsft/if-unofficial-models';


const ccf = new CloudCarbonFootprint();
ccf.configure({
  vendor: 'gcp',
  instance_type: 'n2-standard-2'
})
const results = ccf.execute([
  {
    duration: 3600, // duration institute
    'cpu-util': 10, // CPU usage as a percentage
    timestamp: '2021-01-01T00:00:00Z', // ISO8601 / RFC3339 timestamp
  }
])

Example Impl

The following is an example of how CCF can be invoked using an impl.

name: ccf-demo
description: example impl invoking CCF model
initialize:
  models:
    - name: ccf
      model: CloudCarbonFootprint
      path: '@grnsft/if-unofficial-models'
graph:
  children:
    child:
      pipeline:
        - ccf
      config:
        ccf:
          vendor: aws
          instance_type: m5n.large
      inputs:
        - timestamp: 2023-07-06T00:00 # [KEYWORD] [NO-SUBFIELDS] time when measurement occurred
          duration: 1
          cpu-util: 10

This impl is run using impact-engine using the following command, run from the project root:

npm i -g @grnsft/if
npm i -g @grnsft/if-unofficial-models
impact-engine --impl ./examples/impls/ccf-test.yml --ompl ./examples/ompls/ccf-test.yml

This yields a result that looks like the following (saved to /ompls/ccf-test.yml):

name: ccf-demo
description: example impl invoking CCF model
initialize:
  models:
    - name: ccf
      model: CloudCarbonFootprint
      path: '@grnsft/if-unofficial-models'
graph:
  children:
    front-end:
      pipeline:
        - ccf
      config:
        ccf:
          vendor: aws
          instance_type: m5n.large
      inputs:
        - timestamp: 2023-07-06T00:00
          duration: 1
          cpu: 10
      outputs:
        - timestamp: 2023-07-06T00:00
          duration: 1
          cpu-util: 10
          vendor: aws
          instance_type: m5n.large
          energy: 0.000018845835066981333
          embodied_emissions: 0.02553890791476408