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.

5 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.

Connections 5

Outbound 4
Inbound 1
enables1

Resources 3