Architecture overview

Active-active coordination for multi-node relational databases.

ShardronDB is documented as a coordination layer concept. It is not presented as a replacement for relational databases; it is positioned as an orchestration layer above them.

Traditional HA is often active-passive

Many high-availability database setups protect uptime but leave large parts of infrastructure underused. One node may handle the active write path while others wait as standby nodes.

Single active write pathAll writes flow through one node.

Shared storage dependencyA common storage layer can become a bottleneck or a single point of dependency.

Failover complexityRecovery requires detection, orchestration, and state reconciliation.

ShardronDB concept

The concept introduces a gateway and write coordination layer between applications and active database nodes.

Client / Application
        ↓
ShardronDB Gateway
        ↓
Write Coordination
        ↓
Node 1 Active ↔ Node 2 Active ↔ Node 3 Active
        ↓
Monitoring Panel

Primary key collision prevention

Normal client INSERT requests must not provide PK/id values. Primary keys are generated and coordinated at the system level.

INSERT request
        ↓
Contains id / primary key?
   Yes → Reject
   No  → ShardronDB generates PK
        ↓
Coordinated write flow

Public scope

Layer 1 intentionally avoids publishing runtime source code, replication engine internals, production SQL schemas, deployment scripts, secrets, logs, dumps, or operational credentials.