Document Version

Version Date Author Description
v1.0 27/04/2023 Florent Martinier Initial Document

Introduction

The purpose of this document is to explain all the fields present in an AWS CUR, used by Sopht to calculate energy and financial consumption of AWS services.

Please note that this list may be modified according to updates and new needs from Sopht.

Used Fields

All fields listed below are essential for ecological impact calculations and invoice retrieval.

Field used in CUR Utility
bill/BillingPeriodStartDate Groups usage per month.
lineItem/CurrencyCode Useful for calculating the economic impact in the correct currency.
lineItem/LineItemType Indicates whether consumption is a credit or usage.
lineItem/ProductCode Need to be exhaustive on all products, to map them into larger services.
lineItem/ResourceId
lineItem/UnblendedCost Retrieval of the actual cost charged for the consumption line.
lineItem/UsageAccountId AWS account identification.
lineItem/UsageAmount Quantity of measured consumption.
lineItem/UsageEndDate Groups usage by day.
lineItem/UsageStartDate Groups usage by day.
lineItem/UsageType Need to be exhaustive to know which consumption category we're getting (compute, network, storage).
pricing/unit Adjusts energy calculations based on the unit of measurement.
product/instanceType Retrieves all consumption characteristics of the used instances.
product/productFamily Need to be exhaustive on all products, to map them into larger services.
product/ProductName Need to be exhaustive on all products, to map them into larger services.
product/region Used to adjust consumption based on region emissions.
product/servicecode Used to map services.
product/servicename Used to map services.
product/sku Used to calculate pricing of a service in different regions.
product/transferType
product/usagetype Defines the replication factor of a storage service.
product/vcpu Calculates the energy consumption of a VM.

Data granularity

To perform our analysis, we group the lines according to several criteria:

The data is displayed at the finest level of granularity, which is the day. Quantities spanning several days are proportionally distributed based on their duration during the day.

The start and end of the day are determined in UTC time zone.