The Research Platform is funded directly by your project, with budgets managed by your Project Admin via the RPAdmin console.
- Pay only for what you use: You are charged exclusively for machine runtime and storage.
- Watch your budget: There are currently no automatic shut-offs when funds run low, so it is your responsibility to track your spending.
Funding a Project on the Research Platform
Compute Costs
You only pay for active compute time. Everything on the platform runs in a "container" (a running copy of your software). If a container is running, you pay for it. This includes:
This covers:
| What you start | You pay while… |
|---|---|
| Your notebook workspace | the notebook is running |
| A Dask cluster (many machines for one job) | the cluster is up |
| A background job (a notebook that runs on its own) | the job is running |
How to Control Compute Costs
When you stop a container, the charges stop. The single most important thing you can do to save money is to turn things off when you aren't using them.
- Stop your notebook when you finish for the day.
- Shut down your Dask cluster as soon as the job completes.
Storage Costs
You only pay for the storage space you are actively using across these three areas:
| Storage | What it is |
|---|---|
| EFS | The shared filesystem for your project |
| EBS | Your own private workspace disk |
| S3 | Cloud object storage for large data |
Important Exception: If you connect to S3 storage that sits in a remote AWS account (like your institution's account or a specific mission's account), the Research Platform does not charge you for it. The owner of that remote account pays for that storage, just as they normally would.
Platform Fees and Free Trials
- Platform/Support Fees: None today. We will update this page if a fee is introduced in the future.
- Free Trial: We offer a small free allocation for new projects so you can try the platform. Please ask the Science Cloud support team for the current trial amount when you email them.
What Happens When the Money Runs Out?
Important: Today, the platform does not enforce a hard stop. A workspace will still start even if your funds are depleted, meaning your project can accidentally overspend. You must actively watch your numbers.
How budget tracking works today:
- Platform admins load the project's total budget once the Science Cloud receives the funds.
- Your Project Admin divides that total, setting a specific spending limit for each team member.
- The system prevents the Project Admin from promising more money than the project actually holds.
- Individual members can see their own usage and remaining budget directly inside JupyterLab.
- Project Admins can see the entire team's spending in the RPAdmin console.
What is coming next (Hard Stops):
We are currently building strict budget controls. Once launched, the system will automatically block new workspaces from starting when a project's funds are used up. Project admins will also be able to decide if individuals who hit their personal limits are blocked, or if they can continue spending from the general project pool.
Until this feature goes live, budget control relies on active monitoring and team communication, not system blocks.
Project Funding and RPAdmin Budget Management
Funding and budget management in RPAdmin operate under the following principles:
- Project Accounts: Funding is routed to a dedicated project account, functioning similarly to an AWS linked account.
- Cost Representation: Figures displayed in the RPAdmin dashboard represent AWS service costs, not actual NASA costs.
- Budget Limits: Principal Investigators (PIs) are responsible for setting budgets and spending limits within RPAdmin (similar to AWS Cost Explorer). These limits are used for tracking and should be based on available balance reported by the project report from FinOps.
- Reporting: Until a dedicated self-service dashboard is fully available, projects will receive monthly cost and spending reports to track their allocations and usage.
How to Keep Your Costs Down
- Turn off what you aren't using: Stop your notebooks, Dask clusters, and background jobs the moment you finish with them.
- Pick the right machine size: Choose the smallest size that fits your work. Don't spin up a 64 GB machine for a small task.
- Monitor your budget: Keep a close eye on your personal usage and remaining balance, which are always visible inside JupyterLab.
- Clean up old storage: Delete unnecessary files, models, and data. Stored data incurs a continuous monthly charge.
- Use background jobs for long runs: You don't need to keep an interactive session open just to wait for a long script to finish. Send it to the background so you aren't paying for an active workspace while you wait.