Saving 86% of emissions by moving to the cloud

Project experience | Strategy


Cloudflare, a leading connectivity cloud company based in North America, asked Analysys Mason to quantify the potential carbon reduction that could be achieved by replacing a business’s on-premises network functions with Cloudflare’s cloud-based alternatives. 

We examined a range of enterprise network functions that could be shifted to the cloud, and completed a detailed quantitative analysis of energy usage. We focused in particular on the electricity (‘Scope 2’ emissions) required to power the enterprise network functions, and looked at how consumption would change if produced by an equivalent function hosted in the cloud. Our analysis combined desk research and industry intelligence to analyse data sheets from equipment vendors and information provided by Cloudflare.  

Advantages of cloud-based network functions

Our analysis showed that the largest gain related to optimising the utilisation of hardware. On-premises equipment consumes power constantly but may only be fully utilised for a small part of each day. Shifting that processing work to shared, cloud-based equipment allows far more efficient use of computing resources – cutting energy usage and carbon emissions by around three quarters. 

We also identified a modest but material gain from the efficiency of data centres themselves: modern data-centre equipment makes more efficient use of power (superior power usage effectiveness (PUE)), generating fewer carbon emissions for a given amount of IT output. 

A greener energy mix

The data centres used by Cloudflare are in locations where electricity generation generally has a lower carbon intensity than the global average, which allows a further reduction in carbon emissions relative to an average operator. 

Saving 86% of carbon emissions 

As one of a wide variety of inter-dependent initiatives, cloud-based processing has the potential to make a significant contribution to reducing the TMT industry’s carbon emissions. By moving enterprise network functions to a cloud-based platform, we found that annual carbon emissions from the electricity powering those functions could fall by 86% for a large enterprise (ranging from 78% for a very large enterprise to 96% for a small business).