Operations organizations aren’t always fans of IT. But since 2010, we’ve seen one of IT’s tried and true technology strategies increasingly adopted by operations: virtualization. There’s a good reason for that. They’re seeing powerful benefits from virtualizing their industrial automation (IA) applications.
First, virtualizing Supervisory Control and Data Acquisition (SCADA) and Human Machine Interface (HMI) systems provides huge cost savings by reducing hardware. Virtualization runs multiple applications side by side on one physical server. It also improves server utilization. Instead of 10 to 20% utilization typical in physical models, you can get more like 60 to 70% utilization by virtualizing.
Another advantage is downtime-free upgrades and patches. All you do is create a new VM (virtual machine), load and test the upgrade or patch, and swap out the old VM for the new one in production when everything checks out. This is particularly useful as you adopt new analytics applications, sensor data collection systems, and other elements of the Industrial Internet of Things (IIoT).
You may be wondering: with all my applications sitting on one server, what happens if that server goes down? The weakness with virtualization is that it creates a single point of failure. Instead of losing one application, you would lose them all.
There are several ways to address this problem:
- Run your virtualized applications on a standard server, but maintain a hot or cold standby in case a failure occurs. At best, you’re looking at several hours to get applications back into production; at worst, several days. Either way, expect some data loss.
- Deploy a server cluster and run your virtualized applications in parallel. Clusters offer effective failure recovery, but they introduce complexity and expense. You’ll need two servers, plus switches and additional networking, along with failover scripts. Even with all that, you still risk anywhere from a few to 30 minutes of downtime and potential data loss during failover.
- Get high availability (HA) from a virtualization vendor. Virtualization HA is similar to clusters, although it’s easier to deploy and automate. Regardless, you still have the added expense of another server, with at least several minutes of downtime and data loss while failing over.
- The preferred solution is to prevent failures, downtime, and data loss entirely by putting your virtualized applications on a fault-tolerant system. Here’s a real-world case in point:
A municipal water and wastewater treatment plant runs its virtualized SCADA system on the always-on Stratus ftServer. It’s a single-server solution that operates like any other industry-standard server but is designed specifically for critical applications. This ensures data availability so the plant can satisfy tight EPA regulations. Plus, it lowers costs through reduced hardware, software, and maintenance, helping the municipality cope with a decreasing tax base.
This is a prime example of how you can capture the benefits of virtualization for SCADA/HMI without any of the risks.