Egress considerations
Delta Sharing supports cross-cloud and cross-region data sharing without requiring data replication. However, when data moves across cloud boundaries or between regions, cloud providers charge egress fees to the provider when recipients read shared data. For detailed guidance, see Monitor and manage Delta Sharing egress costs.
Reduce data volume
Partition filtering
Use partitioned table sharing to deliver only relevant subsets of data:
ALTER SHARE my_share
ADD TABLE catalog.schema.sales
PARTITION (territory = 'north_america', year >= 2024);
Change Data Feed
Enable Change Data Feed (CDF) so consumers pull only new or modified records instead of full table refreshes.
History sharing trade-offs
Sharing tables WITH HISTORY enables time travel and CDF, but increases accessible data volume:
- WITH HISTORY: Better performance for incremental reads, but recipients can access historical data
- Without history: Smaller data footprint, but recipients must read full snapshots
Reduce transfer costs
Cloudflare R2
Host or replicate shared data in Cloudflare R2 to eliminate egress fees entirely. Data in R2 can be shared across any cloud or region without outbound transfer charges.
Regional replication
Replicate data to the same cloud and region where your primary customers operate. Use DEEP CLONE with Workflows to automate incremental replication.
VPC endpoints
Encourage recipients to use VPC gateway or interface endpoints for S3 instead of NAT gateways. This reduces costs and enhances security for in-region access.
Monitor and optimize
Use the AWS egress monitoring notebook to identify where costs are incurred and track data transfer patterns. This helps detect high-volume consumers or cross-region traffic driving unexpected costs.
Pattern-specific considerations
| Pattern | Egress impact |
|---|---|
| D2D (same region) | Minimal—data stays within Databricks infrastructure |
| D2D (cross-region) | Standard cloud egress fees apply |
| D2O | Higher—external recipients may not leverage CDF or other optimizations |
Cloud pricing calculators
Estimate your actual egress costs using the official cloud provider calculators:
- AWS Pricing Calculator — S3 data transfer and storage costs
- Azure Pricing Calculator — Bandwidth and data transfer costs
- Google Cloud Pricing Calculator — Network egress and storage costs
- Cloudflare R2 Pricing — Zero egress fees for all outbound data
What's next
- Learn about D2D sharing patterns for Databricks-to-Databricks
- Learn about D2O sharing patterns for external platforms
- Set up monitoring to track sharing activity