Architecture

Active-Active Multi-Site Object Replication

Bidirectional replication between two or more S3-compatible storage sites where all sites accept writes simultaneously, with conflict resolution ensuring eventual convergence.

7 connections 3 resources

Summary

What it is

Bidirectional replication between two or more S3-compatible storage sites where all sites accept writes simultaneously, with conflict resolution ensuring eventual convergence.

Where it fits

Active-active replication enables multi-region or multi-cloud object storage with local write performance at each site. It is the pattern for disaster recovery, data sovereignty, and geo-distributed teams — but adds conflict resolution complexity.

Misconceptions / Traps

Write conflicts are inevitable in active-active. Two sites writing to the same key simultaneously produce a conflict that must be resolved — typically via last-writer-wins, but this can silently drop writes.
Replication lag means eventual consistency. Reads at one site may not reflect recent writes at another. Applications must tolerate or explicitly handle this.

Key Connections

depends_on CRDT — for conflict-free metadata convergence
constrained_by Geo-Replication Conflict / Divergence — the fundamental challenge
scoped_to Geo / Edge Object Storage, Object Storage

Definition

What it is

Bidirectional, conflict-resolving replication between two or more S3-compatible object stores at different geographic sites, enabling read and write access at every site with eventual convergence.

Why it exists

Active-passive replication wastes the passive site's capacity and requires failover during outages. Active-active allows writes at all sites, maximizing utilization and providing true multi-region write capability — at the cost of conflict resolution complexity.

Primary use cases

Multi-region write capability, low-latency global access, disaster recovery with zero RPO, distributed content publishing.

Recent developments

Latest signals

S3 supports two-way (bidirectional) replication rules across regions. AWS native two-way CRR replicates objects in both directions between buckets, with synchronized metadata. The "active-active multi-region" deployment shape is now a single configured replication rule rather than a custom orchestrator. Per AWS Docs — Replicating objects within and across Regions.
Multi-Region Access Points (MRAP) configure two-way replication from one console. S3 MRAP exposes two-way replication setup as a first-class operation — single endpoint serves all regions, traffic auto-routes to the closest healthy bucket. Per AWS — S3 Replication features.
Replication Time Control (RTC) gives near-real-time 15-minute SLA. RTC enforces a 99.99% SLA that replication completes within 15 minutes — closes the "is the failover replica up to date" gap for active-active deployments. Per AWS — S3 Replication features.
One source → multiple destination buckets supported. Same source bucket can fan out to multiple destinations in same or different regions via multiple replication rules — supports topology patterns like "primary + 2 active hot + 1 cold archive" without separate replication pipelines. Per Resilio — Configure & Speed Up S3 Cross-Region Replication.
Conflict resolution: last-writer-wins is AWS default; CRDT semantics require app-layer logic. S3 CRR uses last-writer-wins on conflicting writes to the same key — for true conflict-free convergence on application-meaningful state (counters, sets, document edits), application-layer CRDTs sit on top of the S3 replication substrate. Per Disaster Recovery on AWS — Bi-Directional Cross-Region Replication.
Top 2026 use cases: multi-region live traffic, disaster recovery with zero RPO, distributed content publishing. The active-active pattern's three killer applications — global SaaS apps with regional write endpoints, financial-grade DR (RPO=0), and CDN-style content publishing where every region must serve writes locally. Per AWS for Engineers — Top 5 Use Cases for S3 CRR.