Pain Point

Cross-Region Consistency

The challenge of maintaining a consistent view of S3-stored data across multiple geographic regions when replication introduces latency between writes in one region and visibility in another.

3 connections 3 resources

Summary

What it is

The challenge of maintaining a consistent view of S3-stored data across multiple geographic regions when replication introduces latency between writes in one region and visibility in another.

Where it fits

Cross-region consistency affects multi-region lakehouse architectures where data is written in one region and read in another. S3 Cross-Region Replication is asynchronous by default, meaning a query in the replica region may read stale data or miss recently written files.

Misconceptions / Traps
  • S3 provides strong read-after-write consistency within a single region but makes no cross-region consistency guarantees. CRR replication lag can range from seconds to hours depending on object size and queue depth.
  • Table format metadata (Iceberg metadata.json, Delta _delta_log) must be replicated along with data files. If metadata replicates before data files, queries will fail with file-not-found errors.
  • S3 Replication Time Control (RTC) provides an SLA (99.99% of objects within 15 minutes) but is not a consistency guarantee. It is a best-effort latency bound.
Key Connections
  • scoped_to S3, Object Storage — consistency across S3 regions
  • constrains Active-Active Multi-Site Object Replication — replication lag breaks consistency
  • relates_to Data Residency — multi-region designs must balance residency with consistency
  • amplifies S3 Consistency Model Variance — cross-region adds another consistency dimension

Definition

What it is

The challenge of maintaining consistent views of S3 data across multiple AWS regions or across S3-compatible providers in different geographies, where replication lag introduces windows of inconsistency.

Connections 3

Outbound 3

Resources 3