You are viewing an old version of this page. View the current version.

Compare with Current View Page History

Version 1 Next »

Actual Presented Costs vs. Cloud Estimates

While commercial cloud providers offer cost exploration tools, these displayed costs should be treated as estimates only. Discrepancies occur because native cloud dashboards do not account for:

  • Enterprise support costs
  • Special pricing agreements or negotiated discounts
  • Dynamic pricing adjustments based on volume usage

Relying on these estimates can lead to variances of ~10% (and sometimes more), which could force premature project shutdowns due to falsely predicted budget shortfalls. To prevent this, the Science Cloud provides fully accurate cost reporting by incorporating all actual discounts, refunds, and overhead from official NASA invoices.

How Cost Accuracy is Maintained (The Reporting Lag)

The Science Cloud maintains accurate cost reporting through monthly reconciliation with ARC's official reports. However, because ARC's reports lag by one month, a verification gap exists.

  • Example: When Science Cloud publishes August costs in early September, ARC has only verified data through July. The team relies on estimates for August.
  • Reconciliation: Unverified costs are marked as (estimate) and are reconciled once actual totals are known the following month (e.g., August estimates are verified in October).

Understanding Cloud Account Balances

Cloud account balance reports (funding minus costs) reflect only funds obligated to the cloud contract, excluding overhead expenses (which are obligated separately to the overhead contract).

  • Obligated Funds: Balances are automatically updated when ARC notifies the Science Cloud of new cloud contract obligations.
  • Expected Funds: If a fund transfer misses the mid-month deadline, it appears as an "Expected" line item. This does not count toward your total balance, but serves as a reminder that a funding effort was initiated but not completed.

Detailed Cost Views

Summary views alone don't provide enough detail to understand cost drivers. The FinOps toolkit allows you to view costs by specific categories:

  • Linked Account Name (The specific AWS account)
  • Product/Service Name (e.g., S3, EC2, Lambda)
  • Seller (AWS direct vs. AWS Marketplace)
  • Usage Type (e.g., data transfer, compute hours)
  • Usage Quantity (Volume of resources consumed)

Filtering by Cost Tags: You can filter views to answer specific questions, such as "How much did this user spend on this specific project?"

  • Single tag: View costs for just project1 or project2.
  • Multi-tag: Select user1 and project1 to see all costs for user1's work specifically on project1.
  • No labels