Q&A: Best practices for industrial automation infrastructure upgrades

John Fryer, Stratus Technologies

John Fryer is the senior director of Industry Solutions at Stratus Technologies, where he is responsible for go-to-market strategies and industry initiatives across the company’s product lines. He has more than 25 years of experience with industrial control systems and software products in a variety of engineering, marketing and executive roles at successful startups and major companies, including Motorola, Emerson Network Power and Oracle. His experience includes more than 15 years working with high-availability solutions for the enterprise, industrial automation and networking industries.

On Thursday, Nov. 9, 2017, Fryer will lead a Processing webinar, "Why a modern operational infrastructure is crucial for industrial automation" (1 PM ET for U.S. & Canada, 1 PM GMT for EMEA). During this Stratus-sponsored webcast, attendees will examine the steps that automation organizations can take to prepare for this new digital world and demonstrate how incremental investments can streamline existing operations to generate near-term savings while laying the foundation for future IIoT and Industry 4.0 applications.

The Processing team recently sat down with Fryer to ask a few questions and provide a taste for what will be covered during the webinar.

What are some of the key considerations organizations must be mindful of when upgrading their industrial automation infrastructure?

Modern industrial control systems can be significantly more complex than their predecessors. The increasing emphasis on collecting more data at faster speeds to optimize efficiency, maximize safety and streamline overall business operations requires a more holistic view. Introduce advanced analytics, Industry 4.0 and IIoT into the equation and the thought of upgrading can be daunting.

In this situation, it is important to take a step back and think about what the objectives are from an automation perspective. An upgrade is the chance to address issues that you may have lived with for years such as the complexity of keeping multiple system running, the challenge of upgrading operating systems, various industrial automation applications, keeping HMI’s and thick clients running and up to date and combating cyber security threats.

Some key questions you might want to consider are:

• What is your tolerance for downtime?
• How does control system downtime impact operational efficiency?
• What new applications might you be asked to deploy in the next five years?
• Is it possible to simplify your control infrastructure (and make security easier to manage)?
• What can be done to minimize the reliance on IT?
• Will you need to re-train personnel and require specialized skillsets for an upgraded infrastructure?
• What challenges will there be in migrating from the old system to the new?

What you need to consider obviously differs from plant to plant, but there are many common fundamentals, such as those above, which form the foundation of simple, easily managed and serviced control system infrastructure.

What are some of the key technologies and strategies that enable organizations to realize a modern industrial automation infrastructure?

Upgrading a control system is not a simple proposition. Particularly in process industries, interruptions must be minimized, if they are even tolerable in the first place. The PLCs, valves, sensors, actuators, pumps etc. are often a mix of technologies vary in age and manufacturer, with carefully crafted automation applications that have evolved over time. This makes it critical that the new systems can be deployed as seamlessly as possible. A rip-and-replace is generally not a viable strategy from either a time or cost perspective, so careful planning is required for migration and evolution. This generally starts with upgrading the automation applications and the underlying infrastructure, and it represents the opportunity to consolidate old machines running aging software onto a modern compute platform that is simple to operate, easy to manage and provides proactive notification of any problems. Embracing proven technologies such as virtualization enables the consolidation of many physical machines onto one and provides the flexibility to easily migrate existing applications, install new ones and perform upgrades without complex physical infrastructure configurations. This forms the foundation for the rest of the upgrade cycle and for adopting new future applications.

How does virtualization serve as the foundation of a simplified and cost-effective industrial automation infrastructure?

The traditional approach to implementing automation control systems has been to deploy a single application on a single machine. For many automation systems, this was the only option until about 10 years ago. Virtualization enables all the physical machines to be consolidated onto a single machine. Each individual application now becomes a "virtual machine" and has no knowledge that the other applications (virtual machines) are running on the same physical device. Everything now becomes managed by software. New or updated applications are created with new virtual machines, which can run alongside existing applications. Once a migration is complete, the virtual machine running the old application is simply removed from service and deleted. No physical re-cabling or machine installations are required. Physical infrastructure is greatly reduced, and overall management is simplified, plus deploying new applications becomes much easier.

What are some of the best practices industrial organizations can employ to protect virtual environments?

The main criticism leveled at virtualization is that if the compute platform fails, you lose the entire control system – SCADA, HMIs, historians etc. – and that is unacceptable. The good news is that there are a range of options available to automation engineers to protect against such an event. These include:

• "Regular" swap outs of hardware or a cold standby, i.e. take the risk that the platform won’t fail.
• IT-like solutions involving clustering or virtualization high-availability solutions. These are failure recovery solutions and involve hardware complexity, scripts and multiple software licenses. Hardware failures result in physical un-cabling and re-cabling and reconfiguration.
• Continuous availability solutions based on fault-tolerant compute platforms. These simple solutions prevent failures, keep automation software running and require only a single software license. Additionally, failed components are hot swappable and the exact replacement part is automatically shipped overnight, so there is no downtime.

How can industrial organizations prevent application degradation and failures that can consume time and resources?

Selecting the right solution is key to solving this problem. In many situations there is not an outright failure, just a degradation in performance, or some other unusual behavior. It is important to look for solutions and partners that do not just provide hardware but who also offer tools and internal predictive maintenance capabilities to proactively inform you of problems, or to work with you in isolating failures.

What are the key success factors for a modern industrial automation infrastructure?

There is no substitute for developing a clear and concise plan for an automation upgrade. That being said, clearly understanding the objectives and your tolerances for factors such as downtime, the importance of the data you collect, and ensuring that your new solution, although it may include new, perhaps unfamiliar technologies, results in streamlined operations that maximize your efficiency and deliver a simple solution. After all, swapping an existing complex infrastructure for a new, complex infrastructure does not really gain anything and can result in new long-term headaches.

What real benefits can organizations expect to gain now and going forward by upgrading their industrial automation infrastructure?

Traditional automation infrastructures often have issues that organizations learn to live with. There simply was no alternative at the time. Expansions were complex and generally required downtime to install new applications and check everything out.

A modern automation infrastructure, implemented correctly, provides great flexibility, is simple to operate, is easily serviced and very cost-effective over its lifespan. It frees the automation engineer to work on new projects to optimize operations, not constantly address issues with existing systems.

The automation industry is entering a period of change where new applications, such as advanced analytics in various forms and connectivity to business systems, will drive more rapid software updates and require more integration across an enterprise. Having the right foundation for a control infrastructure can make these future changes easily integrate with existing operations, rather than be a major upheaval to ongoing operations.

The Processing webinar "Why a modern operational infrastructure is crucial for industrial automation" will be presented at 1 p.m. ET, Nov. 9,  for attendees in U.S. and Canada and 1 p.m. GMT for attendees in Europe, Middle East & Africa.

To register for the U.S. and Canada webcast, click here.

To register for EMEA webcast, click here.