system-type-enterprise-integration
Installation
SKILL.md
System Type: Enterprise Integration
Patterns, failure modes, and anti-patterns for connecting, modernizing, and orchestrating enterprise systems.
1. Integration Patterns
Enterprise integration is the art of making systems talk to each other when they were never designed to. Every pattern trades coupling for something — throughput, consistency, simplicity, or debuggability. The right choice depends on what you can afford to lose.
The Topology Spectrum
| Topology | Description | Coupling | Failure domain | Debuggability | When it works |
|---|---|---|---|---|---|
| Point-to-point | Direct connections between each pair of systems. N systems → N×(N-1)/2 connections. | Tightest — every system knows every other system's interface | Localized but unpredictable — one system's outage cascades differently depending on who depends on it | Easy when there are 3 systems, impossible at 15 | Small number of integrations (<5 systems), throwaway prototypes, or when two systems genuinely have a 1:1 relationship that won't grow |
| Hub-and-spoke (ESB) | Central bus mediates all communication. Systems only know the bus. | Moderate — systems decouple from each other but couple to the bus | Centralized — the bus is a SPOF for everything | Good — the bus logs everything (in theory) | When you need protocol translation, message transformation, and routing in one place. When governance requires a central integration team. When the number of systems is moderate (10–30) and the integration team has the capacity. |
| Event mesh / event backbone | Distributed event infrastructure (Kafka, Pulsar, etc.). Systems publish and subscribe to topics. No central broker logic. | Loosest — producers don't know consumers exist | Distributed — broker failure is partial, consumer failure is isolated | Harder — requires distributed tracing, topic monitoring, consumer lag tracking | When you have many (30+) systems, event-driven architectures, and teams that can own their own consumers. When you need to replay events for recovery or new consumers. |