Dataplex Catalog Analytics Hub Cost Sharing

What the solution is

This architecture combines Dataplex Universal Catalog for centralized metadata management and asset discovery with BigQuery sharing (Analytics Hub) listings and Requester Pays Cloud Storage buckets to enforce zero-copy data publishing and subscriber-funded compute.

How it meets the requirements

• Discoverability & Metadata: Dataplex Universal Catalog automatically indexes technical schemas for BigQuery tables and curated data lake filesets while allowing publishers to attach business glossaries, tags, and documentation URLs.
• Cost-Sharing Model: In BigQuery sharing, the publisher pays solely for the persistent storage of the data. Subscribers pay for all compute resources consumed when running queries against their local linked datasets (using on-demand or capacity-based reservations). Similarly, enabling Requester Pays on Cloud Storage buckets ensures that API and egress network charges are billed to the requesting consumer project rather than the publisher.
• Eliminating Replication: BigQuery sharing establishes zero-copy linked datasets in subscriber projects that serve as live read-only pointers to the source data, preventing data drift and redundant storage costs.

Key advantages

• Zero-Copy Architecture: Allows hundreds of external and internal consumers to access real-time data without pipeline orchestration or data synchronization.
• Granular Cost Attribution: Shifts variable query compute and data extraction costs directly to consumer billing accounts.
• Centralized Governance: Standardizes metadata tagging, searchability, and data access policies in place without data movement.
• Secure Publishing: Preserves security controls such as row-level access and column-level security across linked datasets.

Why this is optimal

This approach aligns with Google Cloud data mesh best practices by separating data storage from analytical compute, providing robust discovery, and automating cost attribution across organizational boundaries.