A Pathway to Wisdom in Operational Technology

 

July 27, 2020  |  By Aaron Watkins

Many customers that we speak to are working in the industrial space, some of which have thousands of sensors deployed using operational technology (OT). Many of them consolidate sensor data into a Data Historian system. Such customers have advantages over others using low-cost sensors in that they already control their data; it’s not locked away in a vendor platform.

However, when they want to take advantage of the latest advances in IT, they’re faced with many of the same problems that a low-cost sensor customer has: how do you consolidate data from multiple devices, manufacturers, sites and technology generations, transform it to a standardised form and take advantage of it within your existing business processes?

Thinking in the context of the DIKW Pyramid, businesses are seeking the acquisition of both knowledge and wisdom, but first, must gain access to the raw data and transform it into useful information. This represents one of the core capabilities of Reekoh Accelerate™, with flexible, generic acquisition of data irrespective of source and transformation into a usable form.

A common pattern that we see implemented on the platform is to land data into an unstructured database so that secondary processing (ie. the wisdom and knowledge part) can be performed, before feeding back in for various application integrations.

Reekoh Accelerate™ facilitates ready acquisition of data from Enterprise applications, open data sets, but most particularly, physically generated data. Whether devices have cellular technology and communicate to Reekoh directly; operate on low-power networks and transmit through the network provider; or are legacy sensors sitting on a factory floor, the core platform has a flexible plugin architecture that allows ready integration with these solution components and the ability to expand with new plugins to meet new requirements.

Download our Industrial Edge Integration Architectures

For those systems that have limited connectivity from the edge, we are working on a range of options through our development of Reekoh Outpost™, which should aid data ingestion from the edge. One of the first major steps we’ve made in this area involves our partnership with Skkynet and the publication of some reference architectures that will facilitate MODBUS or OPC based data in a secure manner.

Having acquired that data, invariably, it will not be in quite the form that is desired or may be lacking in standardisation across the range of devices in the field. A factory, winery or building may have technology that has been acquired over the course of a number of years and even those performing a similar function may collect different metrics or at least produce them in different forms.

A recommended approach for firms in this situation is to apply a rules-based decision process before making use of visual mapping capabilities to transform data from its input form into a domain-based Common Data Format (CDF), sometimes known as a Common Message Model, suitable for their business needs.

Whilst introducing a CDF can be complex, it helps to promote interoperability through making it easier to swap out devices (you just need to build an adapter for the new device) and accordingly, helps decouple our solution from the underlying physical devices. In large organisations (or large industries), it can help make resourcing easier. For example, standardisation on the Sparkplug payload format makes it more likely that you can hire integration specialists with pre-existing knowledge of that format and reduce your onboarding times.

Such an approach lends itself to an initial pipeline that may look like the following image:

 

Data ingest and transformation

 

Within Reekoh Accelerate™, you would implement these CDFs as a message schema, which are reusable assets that can be leveraged as your business expands the solution across multiple sites or augments existing datasets with new sensor types. The low-code visual mapping capabilities permit you to largely drag and drop input data segments onto the common format, reducing the level of technical expertise required to perform required data transformations.

The next post in this series will explore approaches that you can utilise in complementary scenarios, leveraging your effort in the first step of gaining wisdom, whilst laying the ground work for minimising effort in your core IT team to support new initiatives.


This article was originally posted on LinkedIn on March 25, 2020.