How Resiliency Drives Cloud Carbon Emissions

Distributing workloads across multiple locations helps users achieve resiliency. Users should be aware, however, that greater duplication can mean both higher costs, and greater carbon emissions. Using public emissions and cost data, this report quantifies the carbon footprint of several common architectures used in building resiliency into stateless cloud applications, and determines the level of protection and cost of each.

Key findings:

• Increasing resiliency by duplicating workloads within a single availability zone costs 45% more than for nonduplicated resources, but nearly doubles the carbon footprint.
  
  
• Duplicating workloads across availability zones, instead of within the same zone, provides higher resiliency — usually without increasing costs or carbon.
  
  
• Greater use of resources (and poor utilization) in active-active configurations significantly increases carbon emissions.
  
  
• Users pay more than double for applications protected against regional failure, but carbon emissions nearly triple.
  
  
• Active-failover configurations provide good protection, with relatively low carbon impacts, at relatively low cost. But recovery times can be long, and capacity not always available.
  
  

Report Authors:

• Dr. Owen Rogers, Research Director for Cloud Computing, Uptime Institute