Why the edge has moved to the forefront

As more manufacturing and process companies discover the tremendous possibilities of pushing intelligence to the edge, expect to see adoption of edge-based industrial control solutions accelerate.

man look at cloud sky sunset
StockSnap (CC0)

Remember just a few years ago, when everyone was talking about cloud computing? While cloud was consuming all the air in the room, few people were paying attention to another technology trend—one with the potential to transform industrial enterprises. I’m talking about edge computing.

The idea of placing computing resources at the network’s edge—at or near where production processes are occurring—is not a completely new idea. Industrial control has relied on distributed computers to control manufacturing machines and processes for decades. But as manufacturers come under increasing competitive pressure, the need to optimize their efficiency, productivity and quality has become a matter of survival. This imperative requirement is driving companies across the industrial spectrum to look at how pushing intelligence out to the edge can help them gain a competitive advantage.

We’ve all seen the images of manufacturing plants with rows of sparkling, computer-controlled robots welding auto body panels or assembling high-tech products. However, the reality is that many industrial enterprises have resisted modernization of their operational technology (OT) environments due to concerns over disrupting existing production processes and capital costs. But that’s changing.

One reason for this change is that these industrial enterprises can no longer ignore the age and inefficiency of their existing control infrastructures. Many processes are controlled by old desktop PCs or servers that are difficult and costly to patch and maintain and may not be secure. Then there’s “server sprawl.” In fact, one brewery we’ve worked with relayed a typical story. With each new application they deployed, a new server was added. Before long, they had no fewer than nine servers—all of which required time and attention to manage and maintain without the benefit of a fully staffed IT organization.

To avoid this, manufacturers are increasingly virtualizing their OT systems, consolidating multiple industrial control applications on a single physical server. This dramatically reduces the cost and complexity of system management, maintenance and security. While virtualization is nearly universal in enterprise data centers, it’s just beginning to pick up momentum in the industrial control infrastructure at the edge. Look for that trend to pick up rapidly in the near future.

A big question now facing the industry is “do cloud technologies have a place in this drive for greater efficiency and agility?” The short answer is that yes, they do – but it will be selective because there are two kinds of application workloads in modern manufacturing and process industry environments: real-time and non-real-time. Real-time applications monitor and measure processes as they occur and take action immediately when intervention is required. These workloads are best performed at the edge, close to where the work is performed to avoid latency issues. Non-real-time applications gather data to help facilitate a range of post-processing corporate analytics. These can be pushed up to the cloud, because latency is not critical for these workloads.

Consider another example from the energy industry. A company may have hundreds of oil drilling rigs dotted across a region. Perhaps hundreds or even thousands of miles away is the company headquarters where the data center/cloud resides. At the edge, on each of the oil rigs, they require systems that provide continuous monitoring and analysis of key parameters, such as well pressure levels, with the ability to immediately identify when critical thresholds are at risk of being exceeded and take immediate action to mitigate them. Waiting for this data to travel back to the data center, undergo analysis and direct actions back to the rig could pose an unreasonable risk. Meanwhile, metrics collected from all of the oil rigs are collected and sent back periodically to the data center/cloud where they can be aggregated and analyzed to support planning and trend-spotting. In this example, both edge and cloud play their most appropriate role.

We’ve seen analysts predict that 40 percent of IoT-created data will be stored, processed, analyzed and acted upon close to, or at the edge of, the network. Process uptime is non-negotiable, so edge systems must stay up and running all the time; but availability is not the only imperative. At the edge, IT skills are either in short supply or non-existent. Consider the oil rig example, or a remote pumping station on a natural gas line; if those systems go down, it could be days or weeks before a technician gets out to service them. So operational simplicity is a must. That means deploying industrial control systems at the edge that are easy to deploy, manage and maintain with the ability to perform remote diagnostics, updating and servicing.

As more manufacturing and process companies discover the tremendous possibilities of pushing intelligence to the edge, expect to see adoption of edge-based industrial control solutions accelerate. Make space, cloud…the edge is coming to the forefront.

This article is published as part of the IDG Contributor Network. Want to Join?

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Now read: Getting grounded in IoT