CASS-Lite: Carbon-Aware Serverless Scheduler on Google Cloud
A sustainability-focused workload scheduler for serverless computing
Hi everyone!
I’m excited to share CASS-Lite, a Carbon-Aware Serverless Scheduler I built on Google Cloud.
It automatically selects the cleanest (lowest CO₂) cloud region at runtime before executing a serverless task.
What Problem Does It Solve?
Carbon intensity varies from region to region, minute to minute.
Developers usually deploy their workloads without considering real-time CO₂ data.
CASS-Lite routes workloads to the greenest region at runtime, reducing carbon emissions without changing application code.
Tech Stack
Google Cloud Functions
Google Cloud Run
Cloud Scheduler
Cloud Logging
BigQuery
ElectricityMaps API
Multiregion architecture
Dashboard for analytics
How It Works
Fetches real-time carbon intensity of 6 global regions
Picks the greenest region
Deploys/executes serverless logic in that region
Logs decisions + carbon metrics
Visualizes impact in a dashboard
Features
Region selection using live carbon intensity
Multi-region Cloud Run triggering
Serverless, pay-per-use design
Logs & dashboards
Extensible to multicloud + Kubernetes
Feedback Welcome!
I’d love to hear:
ideas to extend this into a research project
suggestions for multicloud integration
ways to scale carbon-aware scheduling using Cloud Run/Workflows/GKE
any Google Cloud tools I should explore next
Thanks for reading!