How to get started with systems integration
- Start with a simple integration project like a basic personal data sync
- Begin by identifying the relevant data elements
- Keep security in mind from the start
So you’ve done your research on integration models. Maybe you’ve read my previous posts on integration basics and common terminology. You’ve decided it’s time to change the integration paradigm on campus to use an enterprise data sharing, hub-and-spoke model. And you’re ready to implement a modern integration to make campus more efficient and deliver better services to your end users. Now what? What do you need to do to get ready for an integration?
In a nutshell, you need to understand the business goals of this integration, what it needs to do in support of them, and why it's important. You need to identify what data needs to be exchanged in support of the business needs, what integration capabilities the system(s) you are integrating has, and what you need to build.
With those two key objectives in mind, here’s an overview of the pre-work involved.
1. Define your goals
If you have a digital transformation strategy, articulating the business goals associated with changing your integration within the context of your existing strategy can help get your stakeholders involved and supportive. Define the business processes impacted by the integration and how it helps your end users achieve their goals more efficiently.
2. Choose your first integration project
Deciding where to start can be a challenge itself. Many institutions have dozens of institution-built and maintained point-to-point integrations.
If you’re looking at an existing integration, good candidates include those that would most benefit from real-time access to data. Alternatively, you could look at integrations whose costs weigh on your IT team the most and who may benefit from a more resilient and maintainable approach.
If you’re interested in a simple project to start, it may make sense to start with your basic personal data sync. You no doubt have multiple applications that deal with personal data, so pick one. Why? Because it's easy to understand and it doesn't have a ton of operational complexity—you’d learn how to deliver an integration built with a new set of tools, and since you likely have many systems staying in sync on personal data, you’ll have reusable components when you’re done.
3. Identify the relevant data elements
Once you’ve identified an integration, you need to identify the data it needs. What data will your integrated system need to read, create, or update?
Taking these data elements and mapping them to the enterprise data model will help you understand what APIs your integration will need to utilize.
Because you’re just getting started, building this API list also helps focus your efforts on API enablement, whether that’s the delivered APIs with your ERP or API extensions that bring your custom data.
4. Keep security in mind from the start
Applications that are connected to a hub-and-spoke data-sharing system should get access only to the data they need, through configuration that you control. That data should be exchanged using secure tech that protects data in transit and at rest through the enterprise.
When building your API list in step 3, you are also identifying the data scope your integrated system should be limited to.
5. Get started
APIs that are built on open standards give your development team many tool options for exercising, testing, and ultimately building applications that utilize the APIs.
With your first integration identified, your developers can now start with their tool of choice and begin your integration transformation journey.
Read the other posts in this data integration series: a primer on the whats and whys of systems integration and a guide to common integration terminology.