Traditional enterprise information technology software tools and solutions have found it hard to make inroads into the manufacturing and industrial automation world on the manufacturing shop floor. This can be attributed as the main root cause of IT (Information Technology) and OT (Operational Technology) divide. Machine controllers and data exchange mechanisms have been mostly proprietary to the original machine builders and their automation partners that provide SCADA or HMI for human operators to interact with the machines. This of course has resulted in challenges like poor interoperability, lack of open communication, inability to continually improve control strategies and difficulties in updating or upgrading the software or hardware that runs these machines.

In the IT world as we know the hardware dependent challenges of compute and storage has been resolved over last few decades via software defined compute (a.k.a virtual machines) and software defined storage respectively. This enabled the dawn of Hyperconverged Infrastructure that enabled easy scaling of both compute and storage on-demand. Another advancement we have seen over and above these technologies is the introduction of packaging applications in “Containers”/“Pods” that requires a minimal OS flavor to create a fully functional application engine further reducing the dependence on base operating systems making them operable and run across operating systems. With the advent of containers true interoperability of applications across operating systems is now possible. Public cloud providers like AWS/Google/Microsoft have been providing this infrastructure for quite some time now and this ability of elastic scaling, independence from hardware is the reason for their popularity. However, the adoption of this has been limited to IT tools and system and have not proliferated much on the manufacturing factory floor for OT tools and system. One of the reasons for this can be attributed to low proliferation (virtually non-existent) of Linux operating systems in the OT world. Most of the cloud interoperability, elasticity and hardware independence has been primarily due to deep adoption of Unix or Linux operating system environment. With containerization now it is possible to port OT applications developed on windows environment to run on Linux environments as containers, hence Linux OS is slowly beginning to make its inroads into the OT world.

Another parallel and independent effort that is being made by many automation consortia is towards modular manufacturing controls via MTP (module type packages). The interesting thing is that these MTP packages can be programmed as independent applications that can typically run on a independent container as a microservice. These container applications can then be added to an appropriately sized hyperconverged containerized cluster environment on the factory floor to create a truly plug’n’play, configurable and flexible manufacturing machine operations.

With the availability of inexpensive rugged computing hardware, it is possible to create mini private cloud native Linux edge cluster stacks right on the plant floor to provide same elastic compute and elastic storage provided by public cloud providers and run these OT specific applications as “Containers”. Owing to security, latency and volume of data challenges, the ability to provide this scalable infrastructure as a ‘software + hardware’ combination on-premise near the factory floor is the need of the hour. Another advantage the edge computing clusters provide is the ability to store large amounts of process data right on the plant floor opening the platform for hosting machine learning models to be hosted as containers running alongside the OT applications like MTPs, SCADA or MES applications. A unified messaging environment running on these clusters will serve as a river of data streams acting as a central nervous system that can consume or produce data into or from the data streams river.

A federation of these autonomous micro clusters spread across the enterprise manufacturing sites will be able to provide an ideal way to communicate across machines, enterprise execution systems and perform closed loop manufacturing.