The complete methodology for calculating embodied emissions is:
(((CPU + RAM + SSD + HDD + Motherboard + Power Supply Units + Assembly + Enclosure + Network Storage + Switches) * Disposal) * Time Share * Resource Share)
To summarize the methodology, a worked example for the AWS c6gd.medium instance is provided below. This instance is hosted on the bare metal server c6gd.metal. So, Tailpipe starts by calculating the embodied CO2e of the server hardware components. Tailpipe retrieves the following variable characteristics of the c6gd.metal server:
| CPU name | Graviton 2 |
| CPU units | 1 |
| CPU die size | 457 mm2 |
| RAM units | 8 |
| RAM capacity | 16 GB |
| RAM density | 1.79 GB/cm2 |
| SSD units | 2 |
| SSD capacity | 1900 GB |
| SSD density | 50.6 GB/cm2 |
| HDD units | 0 |
| Network Storage units | 0 |
| PSU units | 2 |
| PSU weight | 2.99 kg |
| Case type | rack |
| Switch Ports | 2 |
Tailpipe inserts these figures into the server GWP calculation (variables in green, constants in blue):
| Component | Formula | Result |
| CPU | 1 * ((457) * 0.0197 + 9.14) | 18.14 |
| RAM | 8 * ((16/1.79) * 2.2 + 5.22) | 199.1 |
| SSD | 2 * ((1900/50.6) * 2.2 + 6.34) | 177.9 |
| HDD | 0 * 31.11 | 0 |
| Motherboard | 66.10 | 66.1 |
| PSU | 2 * (2.99 * 24.3) | 145.31 |
| Assembly | 6.68 | 6.68 |
| Case | 150 | 150 |
| Total | 763.23 |
This means that the lifetime embodied emissions of the c6gb.metal server are 763.23 kg CO2e.
Tailpipe also breaks down the embodied emissions (in kg CO2e) of each component:
| CPU | RAM | SSD | HDD | Others |
| 18.14 | 199.1 | 177.9 | 0 | 368.09 |
Next, Tailpipe calculates the embodied emissions of the instance by factoring in the ratio of instance capability to server capability. First, Tailpipe needs to know the capabilities of both the instance and the server:
| vCPU count | RAM (GB) | SSD (GB) | HDD (GB) | |
| Instance | 1 | 2 | 59 | 0 |
| Server | 64 | 128 | 3800 | 0 |
Then, Tailpipe applies the instance formulae:
| Component | Formula | Result |
| CPU | 18.14 * 1/64 | 0.283 |
| RAM | 199.07 * 2/128 | 3.11 |
| SSD | 177.9 * 59/3800 | 2.76 |
| HDD | 0 * 0/0 | 0 |
| Others | 368.09 * 1/64 | 5.75 |
| Total kgCO2e | 11.9 |
Finally, Tailpipe adds networking-related equipment:
| Component | Formula | Result |
| Instance Components | CPU+RAM+SSD+HDD+Others | 11.9 |
| Network Storage | 0 * 0.0013 | 0 |
| Switch | 2 * 5.06 | 10.12 |
| Total kgCO2e | 22.02 |
This means that the embodied emissions of the manufacture of the c6gb.medium instance are 22.02 kgCO2e.
Next, Tailpipe applies the disposal factor to this figure:
22.02 – (22.02 * 0.018) = 21.62
So, the c6gd.medium instance generates 21.62 kg CO2e of embodied emissions across its entire lifetime.
The customer’s usage of the instance must then be apportioned. To do so, we multiply 21.62 by the customer’s Time Share and by Resource Share:
21.62 * (customer usage period/instance lifetime) * 1
For example, if the customer was utilizing their instance for 18 months over the server lifetime of 6 years:
22.62 * 0.25 * 1 = 5.66
This customer would have generated 5.66 kg CO2e through the embodied emissions of their cloud computing.
That’s the equivalent of driving 21 miles (34 km) in a petrol car.
