Enabling science with the AWS CDK
The AWS Cloud Development Kit (CDK) is something I played with when it was launched in 2019 but was waiting for wider adoption and progression towards feature fullness before trialling again, and it looks like those days have come.
This week I’ve been getting to grips with a developing capability that is aiming to provide radar scientists with a cloud-based platform to generate 2D and 3D radar products for the near future, this is focusing on the 0–2 hour forecasting window and blends observations with forecast model data to generate more accurate output and give Operational Meteorologists a better insight into hard-to-predict convective weather (e.g thunderstorms).
The AWS CDK simplifies how the CI/CD pipelines are described and deployed and also keeps us away from hand-cranked CloudFormation (yippee!). To be able to use control flow logic and modern programming techniques with infrastructure was always a pull of something like Terraform or sceptre, but now we’ve got a fully supported open-source framework and the AWS CDK is the present for some and the future for many.
In combination with GitHub actions, scientists should hopefully be able to concentrate on ‘doing science’ and testing improvements to the science code without having to worry about setting up VPCs and S3 buckets, as we’ll deploy all that for them, as well as put their science code onto an EC2 (although serverless is the goal eventually!) and send a debugging report if needed when they push to GitHub 🚀
If you want to join me on this journey of discovery, make a real difference to society and think you’ve got the skills please apply for the open developer role now! 👨💻
