⚒️ 🎓 Optimize Speed-Critical Workflows Using Parallel Processing (Fan-Out/Fan-In)
⚡ 2,679 views · ⚒️ Engineering
Description
How it works
This template is a hands-on tutorial for one of the most advanced and powerful patterns in n8n: asynchronous parallel processing, also known as the Fan-Out/Fan-In model.
When should you use this? Use this pattern when speed is your top priority and you have multiple independent, long-running tasks. Instead of running them one after another (which is slow), this workflow runs them all at the same time and waits for them all to finish.
We use a Construction Project analogy to explain the architecture:
- The Main Workflow (Top): This is the Project Manager. It defines the project, assigns all the tasks to specialist teams, and then pauses, waiting for a final report.
- The Sub-Workflow (Bottom): This represents the Specialist Teams. It’s a single, reusable workflow that can perform any task it’s assigned.
- Static Data (The Brains): A hidden Project Dashboard is used to track the status of every task in real-time.
The process follows three key phases:
- Fan-Out: The Project Manager starts multiple sub-workflows at once without waiting for them to finish.
- Asynchronous Execution: Each Specialist Team works on its task independently and in parallel. When a team finishes, it updates its status on the Project Dashboard.
- Fan-In: The Project Manager, which has been paused by a
Waitnode, is only resumed when the Project Dashboard confirms that all tasks are complete. It then receives the aggregated results from all the parallel tasks.
Set up steps
Setup time: < 1 minute
This workflow is a self-contained tutorial. The only setup required is to configure the AI model.
- Configure Credentials:
- Go to the
The AI Specialistnode in the sub-workflow (bottom flow). - Select your desired AI credential (Gemini in that case).
- Go to the
- Execute the Workflow:
- Click the “Execute Workflow” button on the
Start Projectnode.
- Click the “Execute Workflow” button on the
- Explore and Learn:
- Follow the execution path to see how the main workflow fans out, and how the sub-workflow is called multiple times.
- Click on each node and read the detailed sticky notes to understand its specific role in this advanced pattern.
🔗 Nodes Used
HTTP Request, Execute Sub-workflow, Execute Workflow Trigger, AI Agent, Google Gemini Chat Model
📥 Import
Download workflow.json and import into n8n:
Workflow menu → Import from File