Healthcare-data platform hardens AWS identity under live incident response

The platform had hit a cost wall familiar to data-heavy SaaS companies: warehouse-resident storage priced for hot analytics, holding data that was queried roughly twice a month. Per-table storage charges accumulated on hundreds of terabytes of largely cold healthcare-claims history — recurring spend delivering no recurring value. Simply switching warehouses would have moved the problem, not solved it. The estate needed an architecture matched to its true access pattern, a safe path to move ~200 TB across clouds without disrupting the SaaS product, and a query layer that kept the data fully analyzable on demand.

VeUP engineered the migration as three deliberate moves. First, transport: a VeUP-built managed data mover carried the estate from Google Cloud Storage into Amazon S3, with schema discovery and mapping translating the Google-side structures to their AWS targets — repeatable, resumable, and verifiable at 200 TB scale. Second, structure: AWS Glue crawlers and the Glue Data Catalog cataloged the landed data, turning object storage into governed, discoverable tables. Third, query: Amazon Athena serves the analytics — selected over Amazon Redshift after VeUP walked the customer through the trade-off, since serverless per-query pricing beats provisioned-warehouse economics for data touched twice a month, while remaining fully capable at multi-billion-row scale. The Amazon RDS-backed web tier serving provider profiles runs alongside the lake, unchanged from the product’s perspective.