Amazon S3 Replication

Amazon S3 Replication is a feature that automatically copies objects from one S3 bucket to another—either in the same region or a different region—for durability, compliance, or performance. Here’s a clear breakdown: 🔁 Types of S3 Replication 1. CRR (Cross-Region Replication) Replicates objects to a bucket in a different AWS region Used for: Disaster recovery Compliance requirements Low-latency access for global users 2. SRR (Same-Region Replication) Replicates objects within the same region Used for: Data aggregation Log processing Compliance (keeping copies in separate accounts) 🔐 Use Cases Disaster recovery across regions Backup and archival strategies Compliance with data residency laws Data sharing across AWS accounts 💰 Pricing Considerations Data transfer costs (especially CRR) PUT requests on destination RTC costs (if enabled) 1. Replication Time Control (RTC) 👉 Guarantees 99.99% of objects replicate within 15 minutes Comes with SLA + monitoring Useful for: Financial systems Compliance / DR (Disaster Recovery) ❗ Extra cost involved ✅ Use it if: You need strict replication timing ❌ Skip if: Normal async replication is fine 🔹 2. Replication Metrics and Notifications 👉 Sends replication status to CloudWatch Track: Pending replication Failed replication Latency Helps in monitoring pipelines ✅ Use it if: Production system / need observability 💡 Recommended for data engineering pipelines 🔹 3. Delete Marker Replication 👉 Replicates delete operations (only for versioned buckets) When you delete an object → a delete marker is created This option ensures delete is also reflected in destination ⚠️ Important: Lifecycle-based deletes are NOT replicated ✅ Use it if: You want source & destination to stay identical ❌ Skip if: Destination is backup (you don’t want deletes copied) 🔹 4. Replica Modification Sync 👉 Syncs changes made to replicated objects Example: Object replicated to destination You modify metadata/tags in destination This option syncs changes back ⚠️ Without this: Changes in destination are NOT tracked ✅ Use it if: Bi-directional consistency required ❌ Skip if: One-way replication (most common case) 🔥 What Should YOU Select? (Best Practice) For most data engineering / production setups: ✅ Replication metrics → ON ✅ Delete marker replication → depends: Backup use case → OFF Mirror use case → ON ❌ RTC → only if strict SLA needed ❌ Replica modification sync → usually OFF