Explanation
Why the waste happens and who it affects.
Much of this traffic is not driven by a resilience or latency requirement: S3 Cross-Region Replication rules that copy whole buckets when only a subset needs protection, AWS Backup copy jobs to a second Region for non-critical data, logs and metrics shipped to a central account in another Region, and applications that call a database, queue, API or bucket that happens to live in a different Region from the compute using it.
These flows are usually set up once, for example by a landing zone, a DR template or a shared service placed in a single home Region, and then grow with data volume. AWS Well-Architected (COST08-BP02) advises avoiding data transfer across Regional boundaries and treating a multi-Region design as a deliberate decision. Relocating individual EC2 instances for regional pricing is covered by the separate Suboptimal Region for EC2 Instance entry; this entry covers replication, backup copy, log shipping and service-to-service flows between Regions.
Billing model
The pricing dimensions that drive this cost.
Inter-Region transfer is charged on the data sent out of the source Region, at rates that vary by service and Region pair.
- Inter-Region data transfer out
- Billed per GB leaving the source Region for another Region, for example $0.02 per GB from US West (Oregon) to US East (N. Virginia) in the S3 pricing example
- Inbound to destination Region
- Recorded as a separate line item with no charge
- S3 Cross-Region Replication
- Adds inter-Region transfer for each destination Region, plus replication PUT requests and destination storage
- AWS Backup cross-Region copies
- Billed for backup data transferred between Regions plus storage of the copy in the destination vault
How to detect
4 checks to find it in your estate.
- In the Cost and Usage Report, filter lineItem/UsageType for values ending in AWS-Out-Bytes (for example USE2-APS3-AWS-Out-Bytes) and group by product and resource to find which services and resources send data between which Region pairs
- List S3 replication configurations and flag Cross-Region Replication rules without prefix or tag filters, or replicating buckets such as logs, build artifacts or scratch data that have no documented recovery requirement
- Review AWS Backup plans for copy actions to other Regions and check whether each copied resource is covered by a disaster recovery requirement
- Use VPC Flow Logs or application dependency maps to find workloads that call databases, caches, queues or APIs in a different Region, and check where shared services and central logging destinations live
How to fix
5 ways to remove the waste.
- Co-locate compute with the data and services it uses most, or deploy regional copies of shared services, so steady-state traffic stays within a Region
- Limit S3 Cross-Region Replication to the data that recovery or latency requirements actually cover by adding prefix or tag filters to replication rules, and remove rules that have no owner
- Keep cross-Region backup copies for workloads whose DR policy requires them, with shorter retention in the secondary Region where allowed
- Aggregate logs and metrics within each Region before shipping, or send only filtered or summarized data to the central Region
- For transfers that must remain, compress and batch payloads to reduce bytes moved, and confirm the design with DR and compliance owners before removing any replication
Documentation
Vendor references for pricing and configuration.