Is IT/OT Convergence Dead?
Don’t panic, I am not implying that the IT and OT organizations are no longer going to be talking to each other. In fact, I think that their collaboration is more important now than before but the conversation that they are having is changing and evolving to account for a much larger trend and this involves the process around how new hardware is introduced into the operations landscape.
IT/OT convergence in the traditional sense that we have always known consists of IT supporting and managing enterprise/business software and OT supporting both a mix of operations hardware and software to manage it. The change and trend that seems to be more and more evident each day, is that IT has become more and more interested in hardware and this is largely driven by the larger market trend of the Internet of Things.
Irrespective of what you think about IoT and its value, it’s my belief that there’s a fundamental flaw in the majority of IoT projects and specifically that these projects involves ignoring the value of the existing OT landscape. A few people may take exception to that statement, and quickly point out connectivity to those systems to pull data out. To this point, I wonder how much time was put into considering how the OT teams will leverage this hardware to help them in their day to day. More often than not, any raw IT device data is immediately pushed upwards to a central (typically cloud) environment and as a result any local opportunity to derive value from it is lost. Another interesting trend is that when IT managed hardware arrives into the physical space of the OT team, there’s a bit of a push-back and as a result diminished value. This is mostly due simply to a lack of understanding of how it can be used locally to support their needs and can lead to both frustration and maintainability issues that ultimately lead to failed projects.
There are two key things in the previously mentioned, approach that are interesting and need to be called out:
- Pushing data upwards can provide value but it’s not differentiating for hardware vendors to achieve this; anyone can do it.
- To be honest, the push of data upwards doesn’t really impress many people and I think that many manufacturing/industrial companies also feel the same way. This is simply because everyone can achieve this and we have seen these scenarios for a very long time…in fact the concept of a “Data Swamp” was already specifically being warned about in 2014, https://www.gartner.com/newsroom/id/2809117, but I think it has been around since the first data warehouses and historians were implemented.
- Local teams see these IT devices/boxes around their facilities, but have no visibility into the data that they are capturing in the context of the rest of the local automation systems.
- The fear of the ‘unknown’ or ‘competition’ for budget and recognition as the experts in the automation space play a role and incorporating the local OT teams is key to building support and leveraging IT managed device data across the industrial landscape (locally and centrally) in the context of the process in which it’s being used.
Diagram 1: SAP Digital Manufacturing strategy to address this Key Industry Trend
What can be done today that will ensure this same issue isn’t repeated, is to setup an enterprise definition of what data is available and an understanding of how it is used to support various industrial processes. This model isn’t just a one-time setup but rather a continuously evolving configuration that mirrors the ever-changing complexities of the business. To achieve this, synchronization at multiple operating levels (Cloud, Enterprise, Fog, & Edge). Additionally, the combination of both automated and manual interfaces is a must so that each level and ensure data accuracy of the structure and the supporting meta-data. This is key to coordinating various systems that are sending data upwards to multiple systems simultaneously.
It is also what we (SAP) have tackled head-on for our customer base, have a look at the latest offerings we have introduced in the SAP Manufacturing Suite and SAP Digital Manufacturing Cloud and for additional details on how to achieve this across the industrial enterprise.
Another related Blog on this topic here around analytics and then persistence of high-volume time-series data is outlined here:
The Inverse Historian (Historian^-1) - https://blogs.sap.com/2018/05/29/the-inverse-historian-historian-1
Hope this helps.