What is the carbon score of your software?
summary
Defines the Software Carbon Intensity as an equation: energy consumed (kWh) × grid carbon intensity (gCO2/kWh) plus embodied hardware carbon, reported per a chosen functional unit (for example carbon per user or carbon per minute). Shows sources for live grid carbon data, cites global average ~519 gCO2/kWh (2019), and notes typical on‑prem server utilization ~10% and data center PUEs ~1.1–1.5. Requires emissions reductions by abatement only, and the SCI alpha ships around COP27.
AI-written summary · about
chapters (23)
- 0:00Intro
- 0:34What makes software green?
- 1:48Greenhouse Gases
- 3:55Is electricity clean?
- 4:34How do we measure electricity?
- 7:28Power usage effectiveness
- 8:29How efficient is an application?
- 9:53How clean or dirty is electricity?
- 11:17Varies by location
- 15:30Sometimes we throw away clean energy
- 16:32Data sources of carbon intensity
- 17:19Energy Transition
- 21:08Hardware is a proxy for carbon
- 23:25Two broad categories of hardware efficiency
- 24:41Extending the lifespan of hardware
- 24:55Increasing the utilization
- 25:46Methods
- 26:51What scope do my emissions fall into?
- 28:10Is it possible to calculate a total?
- 29:31Do totals tell the whole story?
- 30:45Software Carbon Intensity (SCI)
- 32:28Only Eliminations
- 33:55Net-zero strategy
related talks
41mWriting Greener Software Even When You Are Stuck On-Prem
34mSustainable Software Engineering, Building Carbon-Efficient Applications
25mGoogle Cloud Carbon Footprint: Energy-based apportionment [...]
55mToward a carbon-aware Cloud
23mEvaluating the environmental footprint of an open-source software
26mAccelerating Science, Decelerating Carbon: Sustainable Computing...- 25mGoogle Cloud Carbon Footprint: Energy-based apportionment [...]
36mCarbon-Aware Computing: Measuring and Reducing the Carbon Intensity of Software