Integrating an Internet of Things solution can be a challenge, even to seasoned product managers and developers. IoT-connected devices offer so much broad potential that it’s hard to narrow the focus and keep your eyes on the prize. But it’s important to keep some common concerns in mind from the start of the process to the end.
1. Prepare your infrastructure
You need a tool that will accommodate the vast amount of data that is going to be coming from your devices. Normally, it is cheaper and faster to purchase than it is to build, like AWS IoT and ThingWorx.
2. Get a “Logic” layer between your devices and your enterprise systems
In “traditional” integration we’d use an Extract Transform and Load (ETL) paradigm, where data is extracted from one system, business logic transforms the data, and then it’s loaded into a second system. But with IoT there’s no single source system. There are many! Therefore, transformation must either come in line or in batch. The new paradigm is more like Subscribe, Transform, and Publish (STP), where the integration subscribes to a data feed from the device, applies business logic against the incoming data, and then publishes the result to an enterprise system. The tools set for this would be AWS Lambda, the ThingWorx Composer.
3. Have a solid “Device Relationship Management” (DRM) System.
This is like having a CRM for your devices. Like a CRM, a DRM tracks device behaviors, communicates with the devices, understand relationships between devices and systems or even other devices. ThingLogix Foundry was created to fill this need.
4. System Integration Monitoring
Unlike traditional integration where there’s a definite “begin” and “end” time, IoT integrations are continuous, so a robust monitoring and logging system is essential to gain visibility into existing and potential problems.
5. Write your integrations with debugging built in
Normally developers will have a number of log outputs when they are developing integrations to help with debugging. These are normally removed in production. With IoT integration, there are so many more moving parts that it’s difficult to find the exact point of failure. A robust logging system is essential!
6. Plan to scale
Perhaps your solution is small now, but if you’re doing your job correctly, it’ll grow. Make a contingency plan so that you can handle all of your devices and data, even after you grow from 10 connected devices to 1,000 – or even a million! This might involve dedicated teams of device management personnel, a new software solution, or a combination of both.
7. Address security concerns early on
By its very nature, any IoT solution has security concerns because a
ny device that transmits data via a network connection is a potential liability. You know this going in, so build it into the framework of your integration plan. Start by collecting and transmitting only the data that’s necessary, and identifying the data that’s sensitive enough to need extra protection beyond your existing Internet security protocol.
8. Share the wealth
Be a thought leader and a contributor to your industry. Countless people before you pioneered the road that you’re walking now as you integrate IoT connectivity. What did you find valuable? What did you do differently? What did you wish that you’d have known going in? Now’s the time to engage in thoughtful dialogue with other key members of your industry, and share information, just as your IoT devices do.
IoT integration doesn’t have to be overwhelming when you approach it with each of these concerns in mind. The help of an expert consulting team keeps the process flowing as smoothly as possible with an informed perspective on board.