Integration Aggregator vs Point-to-Point
Use an integration aggregator (hub-and-spoke) when you'll connect more than a handful of systems: the hub normalizes data once, so complexity grows linearly. Use point-to-point only for a few stable interfaces. The trap is starting point-to-point for speed and ending in an N-squared web.
Quick answer
More than a few connections, or growing → aggregator (hub-and-spoke). A couple of fixed interfaces → point-to-point. Point-to-point scales as N², and that's the whole story.
Decision table at a glance
| Criterion | Aggregator | Point-to-Point |
|---|---|---|
| Complexity growth | Linear (one spoke per system) | ~N² (interface per pair) |
| Upfront effort | Higher (build the hub) | Lower (one link) |
| Adding a new system | One new spoke | New interface to each consumer |
| Governance & monitoring | Centralized | Scattered per interface |
| Single point of failure | Hub (mitigate with HA) | None central (many small) |
| Best for | Many / growing connections | Few, fixed connections |
What they are
What is Aggregator
An aggregator (hub-and-spoke / integration layer) is a central component that connects to each system once, normalizes to a canonical model (e.g. FHIR), and serves all consumers through one contract. Adding a system means one new spoke, not a new interface for every consumer.
What is Point-to-Point
Point-to-point wires each system directly to every other system that needs it, with a dedicated interface per pair. It is simple and fast for one or two links, with no hub to run, but interfaces grow roughly as the square of the systems.
Key differences
The N-squared problem
Point-to-point is tempting because the first interface is quick. But with M systems that all talk to each other, you trend toward M×(M-1)/2 interfaces, so each new system multiplies the work. An aggregator needs one connection per system, so growth stays linear.
Centralized governance vs scattered logic
A hub gives you one place for transformation rules, mapping, security, audit logging and monitoring. Point-to-point scatters that logic, so changing a data model or compliance control means touching every affected link. For regulated healthcare data, centralization is a real audit advantage.
Upfront cost vs long-run cost
The aggregator costs more to stand up: you're building shared infrastructure before it pays off. Point-to-point is cheaper on day one and more expensive on every day after, once you pass a few connections. Decide based on your trajectory, not just the first interface.
Resilience trade-offs
A hub is a single point of failure you must engineer around with high availability, queuing and back-pressure. Point-to-point has no central chokepoint, but failures are handled per interface with no shared observability. Most teams accept the hub and invest in its resilience.
When to use each
When to use Integration Aggregator
- You connect to more than a handful of systems, or expect to grow.
- Multiple consumers need the same data and you don't want to duplicate interfaces.
- You want one place for transformation, monitoring, security, and governance.
- You're building a product that onboards new health systems over time.
When to use Point-to-Point
- You have a small, stable set of interfaces (two or three) with no growth planned.
- A single direct link is genuinely simpler than standing up and running a hub.
- The integration is a one-off or short-lived and doesn't warrant shared infrastructure.
- Latency or contractual reasons demand a dedicated, unmediated connection.
Frequently Asked Questions
Not sure which fits your build?
Tell us about your product, integration targets, and timeline. We'll map the trade-offs to your context and recommend the path that gets you to a compliant launch fastest.