Allocating Carbon Costs to Computing Payloads across Heterogeneous Infrastructures

As UKRI moves towards a NetZero Digital Research Infrastructure [1, 2] an understanding of how carbon costs of computing infrastructures can be allocated to individual scientific payloads will be required. The IRIS community forms a multi-site heterogeneous infrastructure so is a good testing ground to develop carbon allocation models with wide applicability. Carbon footprints of data centres were measured in the IRISCAST Project [3, 4]. Here we describe and test the models of the IRIS Carbon Mapping Project [5] which allocates carbon costs of compute and storage to individual users and payloads. It is shown that the simple storage and payload models can be implemented and that with improved monitoring enhanced models are also practical.