Reading Time: 7 minutes

A surprise awaited me as I arrived to a recent analyst briefing. My scheduled analyst, the one responsible for integration, had to cancel. His replacement was the analyst responsible for IoT. What were we going to talk about? As it turns out, quite a lot.

From a conceptual perspective, the application network is all-encompassing, supporting large ERP pipes through to tiny sensor readings. As we examine the physicality of the application network, we see that Anypoint Platform is generally concerned with higher-level systems and protocols. Once we go beneath IP, device-specific controllers take over management of these segments. Such controllers will only surface certain information back up to the Anypoint Platform. This type of information flow is of increasing interest in a modern architecture.

Historically, control systems operated independently from enterprise systems. Integrations between Supervisory Control and Data Acquisition (SCADA) or building management systems (BMS), and enterprise systems have been limited to discrete processes, such as asset management and billing. Often these integrations are packed, batched, and leveraged as proprietary middleware.

Service-based delivery models, and the rise of customer 360, increasingly demand the need to combine control and enterprise data. Let’s use an aircraft engine manufacturer as an example. The traditional business model sells fixed assets with a maintenance plan. The modern business model sells operating hours, with hardware and services bundled together. In this context, how can the aircraft engine supplier schedule a customer engine for major service? The answer is to merge machine data with enterprise data in an analytics framework that supports intelligent decision making. This merge is not your good-old Extract Transform and Load (ETL) — it has to happen in real-time.

latest report
Learn why we are the Leaders in API management and iPaaS

The concept that captures this type of system best is that of the ‘Digital Twin.’ A digital twin is a digitized model of a physical thing or process. Through accumulation of data, and the application of analytics and machine learning, a digital twin provides a picture of the past, the present, and potential futures. It enables a platform for actioning and eventing based on the picture.

I am sure there is a marketer out there that is trying to build my digital twin! The insights they derive from my twin will allow them to create personalized experiences for me. This could be real-time data from my wearable devices that show how I am operating, and by marrying this with my product and service history will allow them to dynamically tailor their offerings. Hopefully in a non-creepy way!

To build great digital twins one needs a strong integration muscle. As we have touched on in the discussion above there are a number of integration patterns critical to construction of the twin. Some of the key requirements include:

  • Real-time: The ability to collect large amounts of streamed data is required to capture the current operating environment.
  • In-memory: Will be required to allow for the real-time actioning of signals.
  • Protocol translation: Bridging between wire-level protocols, flow choreography, and semantic formats allows the transmission of information seamlessly between the control and enterprise domains.
  • API-based: APIs to device, APIs to the twin, and between-the-twin APIs for complex systems.
  • Global: From servicing on the shop floor to field operations around the globe.
  • Secure: There will be different views into the twin, engineering will require vital statistics, marketing will require usage statistics. Confidentiality of all data will be a requirement.
  • Scaleable and reliable: A twin supporting mission-critical machinery will need to operate to the highest levels of service necessitating the requirement for high availability and fault tolerance.

Bing! The ring of a bell signaling time-up with the analyst.

I had enjoyed our conversation and we left in agreement that Anypoint Platform had a large role to play in IoT and the build out of digital twins. Through the provision of an integration muscle that supports the major integration patterns Anypoint Platform helps organizations to connect the world’s applications, data, and devices.

Learn more about how Mulesoft is playing in the world of devices.