SA-201b · Module 2

Point-to-Point vs. Hub-and-Spoke

3 min read

The simplest integration pattern is point-to-point: system A calls system B directly. It works perfectly when there are two systems. It works acceptably when there are four. It becomes unmanageable when there are eight, because the number of connections grows quadratically — eight systems with point-to-point connections produce up to twenty-eight integration paths, each of which must be maintained independently.

Do This

  • Use point-to-point for two to three systems with stable, infrequent integration needs
  • Introduce a hub (API gateway, message broker, integration platform) when the system count exceeds four
  • Design the hub as a routing layer, not a logic layer — business logic belongs in the services, not the middleware
  • Monitor the integration surface as a first-class concern — every connection is a potential failure point

Avoid This

  • Start with a hub for two systems — the overhead exceeds the benefit
  • Continue adding point-to-point connections beyond four systems — the maintenance burden grows quadratically
  • Build transformation logic into the hub — it becomes a critical dependency that is difficult to modify or replace
  • Treat integrations as set-and-forget — they require ongoing monitoring and maintenance

Hub-and-spoke centralizes the integration surface through a middleware layer — an API gateway, a message broker, or an integration platform. Every system connects to the hub, not to each other. Adding a ninth system requires one new connection, not eight. The hub pattern trades the complexity of many connections for the complexity of one critical system. Choose which complexity you prefer to manage.