If you knew the state of your machines at each moment, and could take action based on that knowledge, how much value could you create for your customers?
Digital transformation, successfully embedding products and services inside the value chain, creates incredible opportunities for new revenue streams and customer relationships. While embracing new business models created by connected systems can be challenging for enterprises, problems with data management and system integration are more often the cause of industrial IoT project failure. Let’s look at an example ‘ripped from the headlines’ of how decisions today could cause your organization to be left behind.
The harsh reality of industrial IoT is your connected system will be inundated with traffic from many sources and hostile environments. Without centralized data management based on 3 core principles, your IoT data lake will degrade over time from a fountain of knowledge to an expanding tar pit.
Just recently a classic example of the importance of the first principle of identity came to light at Audi, where at least one factory in Germany appears to have produced thousands of vehicles with the same Vehicle Identification Number (VIN) and shipped them to countries across Asia. The VIN is a 17-digit identifier stamped visibly into each car during production, allowing tracking of things like ownership history, maintenance, and accident records throughout the life of a particular vehicle. EU law states a VIN may be used only once every 30 years to prevent overlap. Government investigators discovered the VIN duplication (though not the cause) during their investigation into “Dieselgate,” a scandal prompting regulators in the United States to order Volkswagen (owner of the Audi brand) to recall almost 80,000 vehicles for refunds and repairs. To comply, Volkswagen used the (fortunately) unique VIN of each car to identify and contact their 80,000 unique owners.
Legal problem, solved.
In IoT terms, they used the unique device identifier (VIN) to query for the value (contact info) of a property (current owner) and took action based on the information received (compliance with court order).
System value, delivered.
Now imagine if some of the German Audis shipped to Asia with the duplicate VINs were also subject to the recall. For simplicity, let’s say just one of the cars must be repaired. You enter your query using the VIN of the affected vehicle, and a name appears on your screen… with thousands of other names right below it. Each owner of the thousands of cars with the duplicate VIN is reported to be the owner of the one car you need to find. Now what? That’s a great question that is probably keeping at least one Volkswagen executive up at night.
In their nightmares, they see a Roman general running a simple query to locate the leader of a rebel army in a crowd of defeated soldiers, using his name as the unique identifier. Rather than the system returning a single gladiator for punishment, the executive/general’s screen comes alive with thousands of results all shouting through the speakers, “I Am Spartacus!”
German intrigue and Roman frustration aside, successful business transformation depends on enforcing the unique identification of each connected data source, and maintaining that identity over the life of your system. Even so, data will occasionally enter the system with ambiguous provenance. How will your system adapt? This is not just another ‘feature’ of an IoT application. Identity is a core data management challenge to be solved through proper modeling at the system level, inside the data layer itself. With a trusted relationship between your IoT system of record and your ERP, CRM, and other enterprise systems, data integrity is maintained for enabling intelligent action and creating business value.
Yours and every connected industrial system faces a complex world where sources of data – component parts and sensors – are replaced and swapped between machines. These machines have identities of their own, and are moved between locations and even sold to different organizations. In the midst of this ongoing shell game, manufacturing defects, network errors, and distracted service technicians can cause duplicate or incorrect identity values to enter the system. All of this may have little impact on queries for general trends and distribution insights, where data aggregation can be sufficient. But when you learn that 15% of your pumps and motors installed in customer facilities contain a defective part, proper data management matters. Will you be able to find and replace them all before they go down, taking your customers’ production lines and your business along with them?
If you can’t trust the reported state of your machines at each moment, you can’t take action based on that knowledge. How much value will you create for your customers? Finally a question with an easy answer…