4.9 Assess the impact and requirements of data processing
Hosting, Infrastructure, and Systems
Use carbon-aware scheduling, efficient and secure protocols, and simple system designs to reduce unnecessary processing while balancing performance, security, and constraints.
Success Criteria
- Carbon shifting:
Use carbon-aware computing to dynamically optimize workload execution by scheduling or batching tasks to run at a time when carbon intensity is lower based on real-time grid carbon intensity data, or shifting workloads to regions with lower carbon intensity when possible, while respecting security, performance, and data residency constraints. - Protocol efficiency:
Choose communication protocols that balance user needs, security, and sustainability by improving efficiency and reducing resource use. Prioritize secure options like HTTPS and SFTP/SSH, avoid insecure legacy protocols such as HTTP and FTP, and adopt modern standards like HTTP/2 or HTTP/3 for better performance and lower network overhead. Maintain backward compatibility only where necessary. - Event-driven architecture:
Use event-driven architecture and microservices when building products with state changes that do not require full page refreshes. Favor those where they offer a more energy-efficient alternative to REST-style synchronous APIs based on performance, power, and processing factors, and choose the approach that reduces server workload and environmental impact while meeting requirements. - Data processing efficiency:
Avoid unnecessary data processing. Decide carefully whether processing should happen on the client or server based on performance, security, sustainability metrics, and resource use. - Efficient location:
Perform data transformations, transfers, and processing between the layers of an application as close to the source as possible. This reduces unnecessary serialization overhead and avoids wasting resources.