How do you know what you need from an IoT platform? Use a platform reference model

Written By: Joe Barkai
  • 5/1/2017

When going out to buy a car, people may need to check out the trunk size or the quality of the backup camera, but they know what they’re buying should at least have four wheels, an engine, and some cup holders.

When it comes to buying an IoT platform, that level of experience simply isn’t present. Even sophisticated CIOs and application infrastructure directors aren’t familiar with all of the various enabling technologies that make an IoT solution successful, and most likely are aware that their organization lacks many of the IT competencies required for successful operation of an IoT solution.

How can they measure their needs against the available IoT platforms and determine what is available, what they need to develop or hire for, and how all if it fits together? The problem becomes even more acute when you realize that any solution will inevitably have to mix and match the best offerings from a variety of vendors.

A good way to stay oriented is to use a platform reference model for business solutions.

Understand the structure

An IoT platform is typically distributed across three different levels:

Endpoints: where the sensors and actuators are, and which operate locally, sometimes with a fair amount of local computing power
Gateways: intermediate nodes with more computing power than endpoints, allowing for more sophisticated local management and greater upgrade flexibility
Platform hub: where gateways or endpoints connect, for complex analytics, provisioning, and back-end system integration

Different vendors will have strengths in one or more of these levels, but, given the complexity of even a fairly straightforward IoT implementation, partnership is needed for an effective solution.

Companies generally underestimate the degree of impact an IoT implementation will have on their business applications. It’s essential to recognize the amount of effort it will take to modify existing applications and create new APIs to achieve a new level of function.

Understand the essential IT competencies

Organizations will need to identify and invest to fill competency gaps that will emerge with large-scale IoT implementations. Among skills that might need to be sought are in specifically IoT-related areas of:

  • Distributed multi-tier architecture
  • Event-driven architecture
  • Device management
  • Analytics
  • Digital twins
  • Integration
  • Security

A good example of managing a mix of homegrown and partnered functions, based on a solid IoT platform, is the way Flowserve is monitoring its rugged pumps.

Case study: Flowserve’s Industrial Smart Pumps

Pump manufacturer Flowserve has implemented condition-based IoT asset monitoring of industrial pumps in harsh and remote locations, with various vendors supplying endpoints, gateways, and platform hub. National Instruments supplied sensors and PXI hardware for local data capture. Device management comes from HPE’s Edgeline system. And the ThingWorx platform provides device connectivity, data ingestion, and data analysis, as well as AR-enabled visualization of pump condition. Only a keen appreciation of the competencies required, as well as the choice of underlying platform, has enabled Flowserve to succeed.

See how Flowserve is saving millions with predictive analytics.

  • CAD
  • Retail and Consumer Products
  • Connected Devices

About the Author

Joe Barkai