Build a workload estimate
Count documents, average file size, retained versions, downloads, API traffic, and processing pages. Include OCR, antivirus, AI tokens, indexes, logs, backups, and any external service. Estimate a steady month separately from an initial migration burst.
For example, a collection containing about one terabyte of primary files also generates metadata, requests, and possibly derived outputs. Frequent downloads can cost more than expected from the storage line alone. Archive storage classes have access characteristics and charges to evaluate.
Recalculate for the target deployment
Use the current AWS Pricing Calculator and the relevant regional service prices. Record the assumptions and exclude neither required availability capacity nor operational services from a production estimate.
Review search sizing, network paths, log retention, and processing frequency as explicit decisions. Use budgets and cost allocation to compare actual usage with the estimate after the pilot. Treat documented illustrative totals as examples with assumptions, not standing quotes.