Copyright (c) 2026 MindMesh Academy. All rights reserved. This content is proprietary and may not be reproduced or distributed without permission.
2.4.1.8. Data Cataloging and Cross-Account Data Sharing (Glue Data Catalog, Lake Formation, Redshift Data Sharing, AWS Clean Rooms)
2.4.1.8. Data Cataloging and Cross-Account Data Sharing (Glue Data Catalog, Lake Formation, Redshift Data Sharing, AWS Clean Rooms)
💡 First Principle: Share data by granting access to the one governed copy rather than by copying it, and match the sharing mechanism to the trust between the parties: teams in one organization can see rows they are permitted to see, while separate companies often must see only agreed results.
"AWS Glue Data Catalog": the technical metastore (databases, tables, schemas, partitions) per account and Region that Athena, EMR, Glue and Redshift Spectrum read. Crawlers or ETL jobs keep it current. A business catalog with publish and subscribe workflows sits on top in"Amazon DataZone".- Cross-account sharing with
"AWS Lake Formation": a producer (or central governance) account grants Data Catalog resources to consumer accounts, an organizational unit or the whole Organization, using named resources or LF-tags (tag-based access control scales to thousands of tables). Grants are shared through AWS RAM; the consumer's data lake administrator creates resource links in its own catalog and grants them to local users. Column, row and cell-level filters are enforced for Athena, Redshift Spectrum and EMR, and the data never leaves the producer's S3 buckets. This is the usual data mesh pattern: producing domains own their data, a central account governs access. "Amazon Redshift"data sharing: gives consumer warehouses in other accounts or Regions live, transactionally consistent read access to producer tables with no copies or ETL; each consumer pays for its own compute."AWS Clean Rooms": separate companies analyze their combined datasets without revealing the underlying data to each other. Members join a collaboration and link configured tables that stay in their original location (S3 through the Glue Data Catalog, Athena, Snowflake or Databricks). Analysis rules restrict what can be run: aggregation (only aggregate results, with minimum aggregation thresholds), list (overlap lists of permitted columns) or custom (pre-approved queries and analysis templates). Add-ons include differential privacy, Cryptographic Computing for Clean Rooms (data stays encrypted during queries), Clean Rooms ML lookalike modeling and analysis logs. The collaboration creator designates which member pays for query compute.
| Situation | Mechanism |
|---|---|
| Analytics accounts in one organization need governed access to a central lake | Lake Formation cross-account grants (LF-tags) |
| One Redshift warehouse must read another's live tables | Redshift data sharing |
| Two companies must measure overlap without either seeing the other's rows | AWS Clean Rooms with aggregation or list rules |
| Bulk file exchange with no fine-grained control | S3 bucket policy or access point |
⚠️ Exam Trap: Using Lake Formation cross-account grants (even with row filters) or Redshift data sharing when a partner must never see the other party's records. Those expose the permitted rows to the consumer; AWS Clean Rooms returns only results that the analysis rules allow.
Enroll to unlock the 2 flashcards for this section and review them on a spaced-repetition schedule.
Written byAlvin Varughese
Founder•20 professional certifications