Web apps and data hosting doesn’t always come through the Cloud. Historically, clients have been owning their own machines in their facilities, or in a second time started renting entire machines (CPU) to outsourced suppliers. Then Cloud appeared.
What is the difference between on prem DC, outsourced DC and Cloud services ?
On prem : group of servers that you own and control privately, on your own facilities
Outsourced datacenter : Lease of data center resources from a third-party service provider. Most of the time, the client rents entire machines (physical level)
Cloud : more important than the simple outsourcing of computerized data storage, also provides computing power that can be used remotely via the Internet. Moreover, very flexible because depending on the computer skills, the user can manage their server themselves or just use it remotely in self-service. Most of the time, the client can go in deeper service granularity by the use of virtual machines (a physical machine contains several virtual machines)
Datacenters induce greenhouse gas emissions in several ways:
By their power consumption
equipments and infrastructure
Data transfer (internet network)
By the manufacturing of the used infrastructure and equipments
By refrigerant leakages (refrigerants are used to cool the machines whose energy consumption make temperature increase a lot in DC)
Electrical consumption
👉 Electrical consumption occurs at two levels:
Within the Datacenter
During data transfers with terminals outside the Datacenter
Most of the time, an energy consumption is expressed in kWh ("kilowatt-hours") for facilities' energy consumption.
To carry out this study, Greenly needs:
The Datacenter's power consumption
To be expressed in kWh
For all your sites, or per site
For the total study period, or periodically (monthly for example). If your available data doesn't cover the whole period, our model can extrapolate it
Each site has to be associated to a specific geographical location (country)
If the Datacenter is shared between several users, it is possible to trace the power consumption from one of the following criteria:
Share of your bill in the overall budget of the Datacenter activities
Share of hardware used by your organization
Share of activities related to your organization (In data stored, or calculations made)
If you don't have to total consumed energy consumption but only the consumption for each machine you use, we can determine the whole facility energy consumption by using the PUE ("Power Usage Effectiveness") of your datacenter or use a standard PUE by default.
The amount of data leaving and entering the datacenter
To be expressed in Gigabytes
If this information is not available, it is possible to remove the data transfer from the scope of the study. This will then be specified in your carbon footprint.
The Hardware
👉 Next comes the part related to the Hardware used within the Datacenter.
Manufacturing this equipment and extracting the necessary materials indeed lead to GHG emissions.
To carry out this study, Greenly is based on an IT inventory, which includes this information:
The type of device (server, firewall, switch, etc.)
The model of each device
The manufacturer (Dell, HP, Huawei, etc)
The acquisition date (date at which the equipment got into the DC inventory)
The average lifespan of the types of devices used
If some of the devices used are second hand, you can notify it, and specify how long it was used before arriving in your organization.
In the event that some of this information is not available, the study loses granularity, but Greenly takes care of it by using average data.
Air-conditioned spaces
👉 Finally, air-conditioned area lead to refrigerant gas leakage.
A lot of refrigerant gas are considered as greenhouse gases, just like carbon dioxide or methane.
To take into account these leaks, Greenly is based on :
The refrigerant use
If this information is not available, a reference refrigerant will be used as a basis for calculation.
The amount of refrigerant gas that has leaked during the studied. Two alternatives to determine it :
the total mass of reloaded gas during the study period. This information is most of the time provided by the air conditioning system maintenance operator
leakage amount approximation by using the electricity consumption per site and per refrigerant type