Skip to main content

6.5 Connectivity Patterns

A connectivity pattern describes a recurring integration structure that solves a specific class of problems by combining connectors, datapoints, and (where appropriate) workflows in a consistent way.
Rather than focusing on individual technical implementations, patterns capture the shape of an interaction: which information is collected, how it is combined or evaluated, and where decisions are made.
This makes patterns reusable across domains, technologies, and concrete datapoints.

The purpose of this chapter is not to provide a complete catalog of all possible connectivity patterns.
Instead, it introduces the concept through a single, representative example and shows how the same logical pattern can be implemented using different modeling approaches.
By keeping the pattern itself constant and varying only the execution mechanism, the trade-offs between workflows, connectors, and specialized datapoints become explicit.

In the following sections, the Comparison Pattern is used as a guiding example. At a conceptual level, the pattern is simple: obtain two independent results and compare them in a dedicated step.
What varies is where and how that comparison is performed - directly inside a workflow, delegated to a reusable datapoint, handled by an LLM, or expressed as a compact sequential connector.

By framing integration logic in terms of connectivity patterns, you can reason about solutions at a higher level of abstraction, select the most appropriate orchestration mechanism for a given use case, and evolve implementations over time without changing the underlying intent.

Comparison Pattern​

For a comprehensive guide on the Comparison Pattern, including workflow approaches and connector implementations, see 6.5.1 Comparison Pattern.