Run reconciliation directly from the central Compare environment.
Direct Database Connections keep the topology simple. Compare opens source and target connections itself, which shortens onboarding when network access is already available and the workload stays within a manageable range.
How it works
In direct mode, the Compare service connects to both database endpoints and coordinates the full validation flow centrally. This reduces the amount of moving infrastructure, but it also means the network path between Compare and each database must support the workload end to end.
| Aspect | Direct Database Connections |
|---|---|
| Topology | The main Compare environment connects directly to source and target endpoints. |
| Operational advantage | Fast onboarding, fewer components, and a simpler support model. |
| Main constraint | Higher network traffic and greater exposure to engine-specific sorting differences. |
| Typical fit | Smaller or medium reconciliation scopes in stable, well-connected environments. |
Best-fit scenarios
Fast delivery matters more than deep infrastructure tuning
Use direct mode when the team wants the shortest path from configuration to the first valid reconciliation run.
The Compare host already has reliable access to both databases
Direct mode works best when routing, firewall rules, and latency are already understood and accepted operationally.
The comparison scope stays reasonably compact
Moderate table sizes and predictable workloads are easier to handle without moving compute closer to the databases.
Setup checklist
Confirm endpoint reachability
Make sure the central Compare environment can reach both database hosts and ports.
Configure database credentials
Use accounts with the minimum privileges required for reading and validating the selected scope.
Run connection validation
Check both endpoints with Test Connection before moving into mapping and comparison rules.
Verify sorting-sensitive tables early
Review representative tables when source and target engines differ in collation or ordering behavior.
Move to Agent-Based Connections when row volumes grow, cross-environment latency becomes expensive, or database-specific sorting behavior starts to create avoidable comparison noise.