Steps to Build Your First System Integration

Updated on 02 Jul 2024
2 Minutes to read
Contributors

Article summary

Did you find this summary helpful?

Thank you for your feedback

Prepare and build a system integration with Tulip using best practices to ensure success.

When planning an integration between Tulip and a third-party system, there are a number of and personas to plan with and preparations to settle before you can begin building.

At a minimum, planning should communicate with and involve the following stakeholders:

Citizen Developers
Operations Owners of Integrated Systems
IS/IT Engineering
Shop Floor Users Served by Solution

Together, you can begin considering and making decisions about various aspects of what this integration will involve.

Discuss the problem being addressed.
Establish measurable objectives.
Align on the best solution:
- What is the right data to solve the problem?
- What is the proper System of Record based on the integration needed?
- What IT changes are needed?

Building your First System Integration

Below are high-level steps to guide you on integrating Tulip to a third-party application.
Prerequisite: 3rd Party System must have an API endpoint available (e.g. NetSuite RESTlet, Slack API)

#	Task Description	Who Can Do This?
1	Establish usage of either the Cloud Connector Host or On-Prem Connector Host. (OPCH).	Customer (with Tulip Support if using OPCH)
2	Generate authentication details for Tulip to communicate with 3rd Party System. (e.g. OAuth2.0)	Customer
3	Choose a transaction and decide on the direction of information flow (e.g. GET Work Orders, POST inventory move). We recommend a simple transaction to start.	Customer
4	Expose data and endpoints for that transaction. Document HTTP endpoint and required Inputs (for Netsuite, endpoint = RESTlet with script ID and deploy ID).	Customer
5	Build and test a Tulip Connector.	Customer and/or Tulip (via paid Services project, w/ limited ability to test)
6	Incorporate transaction data into new or existing Tulip Apps with Connector Functions.	Customer and/or Tulip (via paid Services project)

Example Data Flow

Let's say you want to get work orders from an ERP and store those orders into a designated Work Orders table.

First, you need to retrieve the work orders from the ERP. Follow the diagram below for the data flow:

Get Work Orders for today
Integration Diagram Get Work Orders for Today

Then, you need to write the retrieved work orders into a table. Follow the diagram below for the data flow:

Write Work Order to Tulip table
Integration Diagram Write Work Order to Tulip Table

Best Practices

Transact with a source of truth in real-time. In addition, ensure your shop floor is using the most up-to-date source.
Store Tulip-centric data within Tulip. While a workorder’s source of truth may be your ERP, certain data is relevant mainly to Tulip, e.g. non-conformances logged in Tulip against a workorder.
Increase operators' execution of simple ERP/WMS-centric use cases with composable, integrated Tulip apps. For example, a Tulip app using a tablet’s camera as a barcode scanner for inventory management.
Avoid caching data from a source of truth into Tulip tables if the data quickly goes out of date. For example, a Current On-Hand inventory should live in its own datasource, allowing Tulip to interact with it in real-time.
Separately use Tulip for suitable use cases. Any use cases that are best executed in your ERP should remain there.
Use HTTP APIs whenever you can. SQL Connectors require additional configuration.

Steps to Build Your First System Integration

Building your First System Integration

Example Data Flow

Best Practices

Further Reading

What's Next