I was interviewed many years ago for an article about increasing industry curiosity in moving applications to the cloud. At the time the bulk of the discussion was centered around the associated roadblocks and inherent fears. There were not as many pathways for resiliency and redundancy, many applications did not support cloud installations, the commercially available cloud offerings were less diverse than what they were now, and the pathways for compliance in regulated industries were not as defined. This all led to a high resistance to adoption.
A few years after that article was published, I remember sitting in a session for regulatory considerations for validating cloud-based architectures. There was a panel made up of experts and regulatory body representatives from all over the world, and although their opinions, guidance, and anecdotes varied, it was clear a path was being forged. An often talked about but rarely implemented idea was on its way to being more commonly accepted as validation guidelines were being drawn.
Our team began to evaluate edge devices, networking equipment, and example architectures to suss out reliable, repeatable, adaptable, and valid cloud deployments that could be taken to clients for consideration. What came out of this was the realization that there are several different ways to provide resiliency and ensure reliable application access. We canvased our industry sector and learned that each company had different levels of comfort with cloud deployments and varying standards to be followed. It was, however, very clear that cloud deployments moved from something that was not even considered to something that was in bid specs as a viable alternative.
Our foray into cloud deployments started with placing specific software segments in the cloud. A common implementation was having the historized data placed in the cloud along with the reporting platform ensuring local data buffering was happening at the site itself. This method helped to drive an increased desire for cloud deployments of varying scope. As our deployments continued along the same path, focused on data historization and reporting, customers asked us to deploy architectures that could support a complete cloud deployment. This eventually developed into full cloud deployments of specific lines, buildings, or areas as a test of sorts placing the SCADA, data historization and reporting, remote alarm and notification system, and other software pieces in the cloud. Each deployment accomplished redundancy and resiliency differently, but they all provided a framework for future deployments.
Fast forward to now, we’ve had three entirely cloud based deployments in the past six months all located in regulated industries. The architectures relied heavily on edge devices for data buffering and local access in the event of an internet outage and all had several layers of redundancy built on the network side including dual uplinks, ISP redundancy, local application failover, etc. All three of these deployments were requested as the primary option for deployment and not as a viable alternative.
As clients look for ways to decrease overhead costs, real estate requirements, and staffing loads, many manufacturers are turning to cloud deployments to help accomplish this. It is apparent that thanks to technological advances and acceptance of newer IT tools focused on cloud deployments that the scales have tipped from risk aversion to widespread adoption. As more best practices come about, the appetite for cloud deployments will only increase. It is now our job as system integrators to continue to look for ways to protect the enterprise, prove out deployment options, and provide our clients with competitive advantages in their fields.